阴阳 五行 块之 一: 根基 济灵 f 月 之 f 月
MEMORIZABLE PUBLIC-KEY CRYPTOGRAPHY
(MePKC) & ITS APPLICATIONS
1111111
1166611
1611161
1166611
1116611
1116611
1116611
1116611
1666611
1116661
1666611
1111111
Mainly authored by Kok-Wah LEE (李 国华) @ Xpree Li, plus
some ASCII arts of 2D key authored by Wei-Dong Chui (徐 伟栋) & Wei-Jian Chui (徐 伟坚)
Unpublished Copyright © since the First Draft on 25 October 2008 in Malaysia
Publication Copyright © since the First Edition (version 1.0) on 14 March 2009 in the USA
Copyright Statement
The copyright of this book (aka "thesis" due to its role as a collection of
statements and theories put forward and supported by arguments) belongs to the
authors under the terms of the primary US Copyright Act in which it is firstly
published online from a computer server located in the United States of America,
secondly Malaysia Copyright Act 1987 due to author residency and citizenship, as
well as tertiary Berne Convention for all its signatory countries for any international
copyright entitlement.
Due to the relatively high research costs invested by the author, for refund, as
well as for building up a fund for further maintenance, research and development,
any original and novel idea conceptions from the author in this book is only free of
usage for public interests, press report, private study, research, and teaching
throughout the World, with the condition that proper originality citation for source
references has been clearly shown. Yet for any commercial usage, prior consent has
to be obtained from the author or his successor(s).
© Kok-Wah Lee @ Xpree Jinhua Li, 14 March 2009
Email: E96LKW@hotmail.com
URL (h): www.geocities.com/xpree/
URL (o): www.xpreeli.com
All rights reserved
ii
DECLARATION
I hereby declare that the original and novel idea conceptions in this book of research
essays towards a future possible doctorate thesis in information engineering,
generally, and information security, particularly, have all been done by me, except
Appendix B to have some new and creative child-made ASCII arts by Wei-Dong
Chui and Wei-Jian Chui.
李国华 @ 李锦华
Kok-Wah Lee @ Xpree Jinhua Li
Find me Xpree or Xpreeli in the Internet!
Email (Home): E96LKW@hotmail.com
Email (Business): contact@xpreeli.com
URL (Home): www.geocities.com/xpree/
URL (Business): www.xpreeli.com
First unpublished draft on 25 October 2008 in Malaysia
First edition (version 1.0) of online publication on 14 March 2009 in the USA
in
ACKNOWLEDGEMENTS
I hereby would like to express my gratitude to the following persons together with
their inputs that have been given me in completing the electronic book of this PhD
research project, which has contributed mainly in the novel knowledge field of
key/password security leading to the memorizable public-key cryptography
(MePKC) and its applications. Here are the listees:
(i) My parents, relatives and friends together with those anonymous people
- For help and giving me the physical, emotional and spiritual supports.
(ii) Lake-Tee Khaw, University of Malaya, Kuala Lumpur, Malaysia
- For giving the advantages and disadvantages of SD (Statutory Declaration).
(iii) Gita Radhakrishna, Multimedia University, Melaka, Malaysia
- For supplying the legal contents about copyright.
(iv) Alan Wee-Chiat Tan, Multimedia University, Melaka, Malaysia
- For supplying C++ class of big number arithmetic after my given idea; and
- Being a nominal supervisor since lb August 2008 till now.
(v) Voon-Chet Koo， Multimedia University, Melaka, Malaysia
- For prototyping the RJ45 switch using conventional DIP switch.
(vi) Hong-Tat Ewe, Multimedia University, Cyberjaya, Malaysia a
- For being a nominal supervisor since 27 May 2004 till 15 August 2008; and
- For having lots of unpleasant and trust-less interactions.
(vii) Matt Bishop, University of California at Davis, CA, USA
- For supplying a proceedings paper.
(viii) Alex X. Liu, Michigan State University, MI, USA
- For supplying a journal paper.
(ix) Chee-Onn Chow, University of Malaya, Kuala Lumpur, Malaysia
- For supplying a journal paper.
iv
DEDICATION
特将这 本 研究 论文书 献给于 我 敬 爱的父 亲李厚 芳和母 亲徐亚 妹。
This oook or research essavs is dedicated to my respected and beiovea parents,
Hew-Fong Lee and Ah-Mooi Choi.
H-( A _ A )-H
Find me Xpree or XpreeLi in the Internet!
CN: 语言与 文字 是 了解 ー个 文化的 终极 密码。
EN: Language and Writing are ultimate keys to understand a culture.
MY: Bahasa dan Tulisan adalah kunci terdasar to memahami satu ketamadunan.
V
ABSTRACT
The acquisition of a doctorate research degree is qualified by the sufficient and
novel knowledge contribution from a PhD postgraduate student. Generally, the
contribution of a scientific researcher is based on the publication counting, credit of
journal and proceedings, citation counting, qualified peer evaluator, and contribution
impact. Lately, D. L. Parnas (2007) called for a more accurate evaluation of
researcher's contribution instead of the publication counting and citation counting.
Parnas suggested that a researcher should provide some of one's main papers to be
evaluated by some qualified peer evaluators. Here, proposal is delivered to improve
Parnas' suggestion, where the researcher shall also list out the impacts of one's
selected papers like the successful implementation and realization of a research result
for public usage and application. This step of emphasizing researcher evaluation based
on contribution impact is crucial to the welfare of human civilization to avoid the
number game of discussing a war on some plain papers. Subsequently, this PhD
project is carried out based on this evaluation principle of contribution impact, where
the prototypes are expected to be realized and/or commercialized for public usages.
The main knowledge contributions of this research project are some key
management techniques to create big and yet memorizable secret(s), especially 2D
key, memorizable public-key cryptography (MePKC), and their applications in
information engineering. Password the secret is the most popular, easily compatible,
and cost-effective computer authentication method. In the current prior art, a
memorizable secret reaches a maximum size at 128 bits (Schneier, 2006). This has
limited the AES (Advanced Encryption Standard) key at 128 bits and storage
technology of private key to be in the forms of encrypted private key, split private key
(Ganesan, 1996, September 17)， and roaming private key (Baltzley, 2000， November
28).
Here, the invented big secret creation methods and systems are (i) self-created
signature-like Han character of CLPW (Chinese Language Password) and CLPP
vi
(Chinese Language Passphrase), (ii) two-dimensional key (2D key), (iii) multilingual
key, (iv) multi-tier geo-image key, (v) multi-factor multimedia key using software
token, and their hybrid combinations. From every one of these inventions, the AES-
256 key can be realized as well as the other applications in information engineering
needing a big secret. Yet the most important contribution is the realization of fully
memorizable private key towards the MePKC using finite field cryptography (FFC),
elliptic curve cryptography (ECC), hyperelliptic curve cryptography (HECC), or
torus-based cryptography (TBC). The explanation here adopts 2D key and
multilingual key to create a fully memorizable private key for ECC. To resist the
demand for longer key size due to the advancement of computing technologies, key
strengthening is used.
Multimedia semantic noises are studied. Textual semantic noises like
capitalization, punctuation marks, misspelling, mnemonic substitution, permutation,
character stuffing, and ASCII mutual substitution table, are proposed to increase the
randomness of CLPW, CLPP, and 2D key. The available styles of 2D key are
multiline passphrase, crossword, ASCII art, Unicode art, etc.
Due to the technical and legal factors, there are different asymmetric key pairs
for different cryptographic schemes. Hence, there are many private keys to be
supported. From Forrester Research (Kanaley, 2001)， an active Internet user manages
an average of 15 keys on a daily basis. However, Adams and Sasse (1999) reported
that users could only be expected to cope with a maximum of 4 or 5 keys that are
unrelated and regularly used. If one key is used for two or more accounts, there is a
risk of domino effect (Ives, Walsh & Schneider, 2004). Hence, the key management is
a balance issue of usability and security. There exist some methods to generate
multiple slave keys (aka site key) from one master key for multiple online accounts by
using the domain name, random number, single sign-on server, ana key strengthening.
Here, some new key management methods are proposed to extend the generation of
slave keys from one master key to both the online and offline accounts, which is called
multihash key, by using domain name, key strengthening, and hash truncation. A
Vll
brother of multihash key called multihash signature, using multiple hash values of a
message from different hash iteration, is proposed to provide object-designated
signature function. The object may be recipient, action, feature, function, meaning,
etc., as a representation.
Annually, the World and USA lose USD$600 billion and USD$250 billion,
respectively, due to economic espionage (Burgess & Power, 2008). The MePKC
secures the computer communications of an electronic file over an insecure channel
and an encrypted local file in the computer. However, only the ciphertext is protected
and not the decrypted plaintext. Sometimes, the acquisition of many ciphertexts by an
attacker through the pervasive hacking can be in favour of the attacker for ciphertext-
only attack, which is especially effective for the cracking of digital signature. The
current technique to resist hacking by using firewall is effective at the server side due
to experienced network administrator, but not at the client side due to the technical
difficulties and financial cost. Here, a simple method and device using DIP (Dual In-
Line Package) switch is proposed to secure a hacking-free data storage for common
end users.
MePKC can be applied to digital timestamping scheme. Humans are needed to
hold the private key in the form of fully memorizable secret and not the computing
devices. The legal case of laser inventor Gordon Gould (Taylor, 2002)， which lasts for
30 years, has informed us the weaknesses of IP proving method using SD (Statutory
Declaration). Here, MePKC timestamping scheme can be used as the evidence for
general IP (Intellectual Property) like patent and copyright. This is especially effective
for the first-to-invent patent filing system. Today, all the countries in the world are
using first-to-file patent filing system, except the US is still using the first-to-invent
patent filing system. To resist the hacking threats, a limited first-to-invent patent filing
system is proposed. It has a window filing period of one year from the claim of a
timestamp, and uses the timestamp to prove the IP originator(s). Ruth Taplin (2004)
reported that IP assets in 2004 accounted for up to 70% and 40% of market values of
all corporate assets in the US and Japan, respectively. Tms reflects the truth that the
vm
developed countries are now relies heavily on the knowledge economy, where IP is
the main pillar. It also tells the crucial importance to apply MePKC timestamping
scheme.
As a whole, the current and other applications of big and yet memorizable
secret are as follows:
(i) methods and systems to realize memorizable symmetric key the secret
till resistance to quantum computer attack;
(ii) methods and systems to realize memorizable public-key cryptography
(MePKC);
(iii) methods and systems to improve security strength of other
cryptographic, information-hiding, and non-cryptographic applications
of secret beyond 128 bits;
(iv) method and system to harden the identification of embedded data in
steganography although stego-data has been detected;
(v) method and system to transfer fund electronically over a remote
network using MePKC;
(vi) method and system to license software electronically over a remote
network using MePKC;
(vii) methods and systems to authenticate human-computer and human-
human communications at a local station or over a remote network
using MePKC;
^vm) method and system to use digital certificate with more than one
asymmetric key pair for different protection periods and password
throttling;
(ix) method and system to use three-tier MePKC digital certificates for
ladder authentication;
(x) method and system to store, manage, and download voice and video
calls of mobile phone and wired phone at online distributed servers;
(xi) method and system of multipartite electronic commerce transactions;
as well as
ix
(xii) method and system to boost up the trust level of MePKC digital
certificate by using more than one certification authority (CA) and/or
introducer of trust of web.
In a nutshell, some new key management techniques are invented. These
techniques are used to create human-memorizable big secret(s) and further to realize
the MePKC and its applications. The secure hacking-free DIP switch and the
proposed limited first-to-invent patent filing system are hoped to be adopted
prevailingly and pervasively to resist the economic espionage through virtual hacking.
In the long run, the MePKC is expected to be fully utilized to create and promote an
environment-friendly and electronically networked info-computer era as for the e-
commerce communication to be upgraded from bipartite into multipartite
crypto system.
100
Figure 0.1a Multiline passphrase
linn
111111
— — li—
—li—
—li—
111111
linn
102
Figure 0.1c ASCII art
HAPPY ᅮ
M^INCH
SPELLS
ᄉ \ 껴
101
figure 0. lb Crossword
¥¥¥¥¥
©©¥©©
©©¥©©
©©¥©©
¥¥¥¥¥
103
figure O.ld Unicode art
Figure 0.1 Two-dimensional (2D) key
Keywords: Key/password security, big secret(s) creation methods, 2D key,
memorizable public-key cryptography (MePKC), multihash key, multihash signature.
X
TABLE OF CONTENTS
TITLE i
COPYRIGHT STATEMENT ii
DECLARATION iii
ACKNOWLEDGEMENTS iv
DEDICATION v
ABSTRACT vi
TABLE OF CONTENTS xi
LIST OF TABLES xviii
LIST OF FIGURES xx
PREFACE XXV
CHAPTER 1: OVERVIEW 1
1 . 1 Introduction 1
1.2 Motivation 1
1.3 Research Aims 3
1.4 Research Methodology 4
1.5 Organisation of the Thesis 4
CHAPTER 2: EVALUATION ON KNOWLEDGE CONTRIBUTION 7
2.1 Life of Research Postgraduate Students 7
2.2 General Conditions to Get a Doctoral Degree 7
2.3 Prior Approaches of Research Evaluation 8
2.4 Contribution Impact 9
2.5 Research Quality 12
CHAPTER 3: CREATING BIG MEMORIZABLE SECRET (PART 1) 17
3. 1 Required Protection Periods and Their Key Sizes 17
3.2 Review of the Secret for Symmetric Key Crypto system 20
3.2. 1 Related Work: Single-Line Key/Password Field 22
xi
3.3 Review of the Secret for Asymmetric Key Crypto system 22
3.4 Potential Methods to Create Big and Yet Memorizable Secret 25
3.5 Methods and Systems to Create Big Memorizable Secret 27
3.6 Potential Applications of Available Big Memorizable Secret 27
3.7 Future Development of Keys the Secret 29
CHAPTER 4: CREATING BIG MEMORIZABLE SECRET (PART 2) 31
4. 1 Passphrase Generation Methods 3 1
4.1.1 Acronym 3 1
4. 1.2 Full Sentence 32
4.1.3 Diceware 32
4.1.4Coinware 33
4. 1 .5 Monolingual, Bilingual, and Multilingual Word Lists 33
4.1.6 Key Length Requirements of Coinware 34
4.2 Chinese-Character-Encoded Passphrase 36
4.2. 1 Environ Password 37
4.2.2 Unicode 37
4.2.3 Chinese Language Password (CLPW) 38
4.2.4 Key Length Requirements for Coinware in Chinese Language 41
4.2.5 Self-Created Signature-like Han Character 41
4.2.6 Self-Created Chinese Language Passphrase (CLPP) 42
4.2.7 Cantonese Language Password Using Jyutping 43
4.2.8 Japanese Language Password Using Romaji 45
4.2.9 Key Length Requirements for Various Chinese Key Spaces 46
4.2.10 Example to Create CLPW and CLPP ^ 47
4.3 Two-Dimensional (2D) Key 50
4.3.1 Related Work: Single-Line Key/Password Field 53
4.3.2 Related Work: Key Strengthening 53
4.3.3 2D Key Input Method ᄂ ᄂ 54
4.3.4 Styles of 2D Key: Multiline Passphrase 57
4.3.5 Styles of 2D Key: Crossword 57
4.3.6 Styles of 2D Key: ASCII Art I Unicode Art 58
4.3.7 Styles of 2D Key: Colourful Text 59
4.3.8 Styles of 2D Key: Sensitive Input Sequence 59
4.3.9 Applications for Symmetric and Asymmetric Key Crypto systems 60
4.3.10 Conclusion 61
CHAPTER 5: CREATING BIG MEMORIZABLE SECRET (PART 3) 62
5.1 Multilingual Key 62
5.1.1 Related Works 62
5.1.2 Black-and-White Multilingual Key 62
5.1.3 Grid Partitioning for Higher Entropy and Randomness 64
5.1.4 Colourful Multilingual Key " 67
5.1.5 Font Technologies to Solve Data Size Problem 67
Xll
5.1.6 Resistance Techniques to Shoulder-Surfing Attack 69
5.1.7 Fast Search for a Unicode Graphic Symbol 69
5.1.8 Applications for Symmetric and Asymmetric Key Crypto systems 71
5.2 Multi-Tier Geo-Image Key 72
5.3 Multi-Factor Multimedia Key Using Software Token 74
5.3.1 Related Work ' ᄂ 77
5.3.2 Using Key the Secret and USB Flash Drive 77
5.3.3 MePKC Application with Optional Split Key Method 77
5.3.4 Software Prototype 78
5.4 Hybrid Combinations 78
CHAPTER 6: MULTIMEDIA NOISES FOR MORE RANDOM SECRET 79
6.1 Introduction 79
6.2 Semantic Error Occurrences in the Multimedia Communications 79
6.2.1 Introduction 80
6.2.2 Multimedia Type of Bit Stream 81
6.2.3 Multimedia Type of Text 81
6.2.4 Multimedia Type of Image 85
6.2.5 Multimedia Type of Audio 86
6.2.6 Multimedia Type of Video 88
6.2.7 Duplication of Multimedia Data 89
6.2.8 Conclusion 89
6.3 Decrypting English Text Using Enhanced Frequency Analysis 90
6.3.1 Introduction 91
6.3.2 Justification on Frequency Analysis 91
6.3.3 Drawbacks or Conventional Frequency Analysis 94
6.3.4 Combined Methods of Frequency Analysis 95
6.3.5 Proposed Method: Enhanced Frequency Analysis 96
6.3.6 Conclusion 98
6.4 Passphrase with Semantic Noises and a Proof on Its Higher Information
Rate ᄂ 99
6.4.1 Introduction 100
6.4.2 Key Sizes and Passphrases 101
6.4.3 Unicity Distance and Passphrases with Semantic Noises 102
6.4.4 Proof of Higher Information Rate 106
6.4.5 Conclusion 107
6.5 Information Rates and Unicity Distances 108
CHAPTER 7: MULTIHASH KEY 112
7. 1 Introduction 112
7.2 Multiple hashes of single key with passcode for multiple accounts 1 12
7.2.1 Introduction 113
7.2.2 Related Works 116
7.2.3 Basic Model ofMultihash Key 118
Xlll
7.2.4 Discussions 124
7.2.5 Limitations 126
7.2.6 Conclusion 128
7.3 More Slave Keys for Offline Accounts per Master Key of Multihash Key 128
7.3.1 Related Works 131
7.3.2 First Method: Using Filename 132
7.3.3 Second Method: Using a Random Number 133
7.3.4 Third Method: Using a Two-Tier Structure 135
7.3.5 Conclusion 136
7.4 First Variant of Multihash Key Using Automatically Selected Tiers and
Security Levels 136
7.5 Second Variant of Multihash Key Using Automatically Selected
Permutation Sequence of Security Levels 138
7.6 Third Variant of Multihash Key Using Hybrid Combination 141
7.7 Fourth Variant of Multihash Key as a Further Authentication Factor 143
7.8 Fifth Variant of Multihash Key as a Simple Key Escrow Method and
System 145
7.9 Discussions on Variants of Multihash Key 146
CHAPTER 8: MULTIHASH SIGNATURE 147
8.1 Introduction 147
8.2 Classical Digital Signature Scheme 149
8.3 Applications of Hashing 149
8.4 Proposed Method: Single Message with Multiple Digital Signatures from
Single Asymmetric Key Pair 150
8.5 Comparisons for Advantages 152
8.6 Conclusion 156
CHAPTER 9: APPLICATIONS OF BIG SECRET & MePKC (PART 1) 157
9. 1 Applications of Created Big Memorizable Secret(s) 157
9.2 Memorizable Symmetric Key to Resist Quantum Computer Attack 158
9.3 Memorizable Public-Key Cryptography (MePKC) 159
9.3.1 Related Works: Storages of Private Key 159
9.3.2 The Proposed MePKC Appliacation 9.1(ii) 160
9.4 Other Cryptograpnic, Information-Hiding, and Non-Cryptograpnic
Applications of Secret beyond 128 bits 163
CHAPTER 10: APPLICATIONS OF BIG SECRET & MePKC (PART 2) 165
10.1 Identification Hardening of Embedded Data in Steganography 165
10.1.1 Related Works 165
10.1.2 The Proposed MePKC Appliacation 9.1(iv) 165
10.2 Electronic Fund Transfer Using MePKC 168
xiv
10.2.1 Related Works 168
10.2.2 The Proposed MePKC Appliacation 9.1(v) 172
10.3 Electronic Software Licensing Using MePKC 172
10.3.1 Related Works ᄂ ᄂ 172
10.3.2 The Proposed MePKC Appliacation 9.1(vi) 173
CHAPTER 11: APPLICATIONS OF BIG SECRET & MePKC (PART 3) 177
11.1 MePKC Human-Computer and Human-Human Authentication Schemes 177
11.1.1 Related Works: Computer Password Authentication Protocol 1 77
11.1.2 First Model without Perfect Forward Secrecy 179
11.1.3 Second Model with Perfect Forward Secrecy 1 80
11.1.4 Re-Authentication Rules 185
11.2 MePKC Digital Certificate Having More Than One Asymmetric Key Pair
for Different Protection Periods and Password Throttling 1 86
11.2.1 Related Works: Digital Certificate and Password Throttling 1 86
1 1 .2.2 MePKC Digital Certificate with Many Asymmetric Key Pairs 1 87
11.3 Three-Tier MePKC Digital Certificates for Ladder Authentication 1 89
11.3.1 Related Works: Digital Certificate and Ladder Authentication 1 89
11.3.2 Three-Tier MePKC Digital Certificates for Various Applications 1 90
11.4 Anti-Phishing Using MePKC Authentication Schemes 193
CHAPTER 12: APPLICATIONS OF BIG SECRET & MePKC (PART 4) 194
12.1 Archiving the Voice/Video Calls of Wired/Wireless Phones 194
12.1.1 Related Works 194
12.1.2 The Proposed Model 195
12.2 Multipartite Electronic Commerce Transactions Using MePKC 197
12.2.1 Related Works ᄂ 197
12.2.2 Artificial Neural Network Based Byzantine Agreement Protocol
(ANN Based BAP) ᄂ 199
12.2.3 Entity Partitioning in the Electronic Commerce Transactions 200
12.2.4 Tripartite ANN Based BAP with Trusted Party 207
12.2.5 Tripartite ANN Based BAP without Trusted Party 207
12.3 Trust Boosting of MePKC Digital Certificate by Using More than One
Certification Authority and/or Introducer of Trust of Web 207
12.3. 1 Related Works: Risks of Public Key Infrastructure 207
12.3.2 Some Human Interaction Models 209
12.3.3 Model to Boost up the Trust of MePKC Digital Certificate 218
CHAPTER 13: MePKC TIMESTAMPING SCHEME FOR EVIDENCE
OF INTELLECTUAL PROPERTY (IP) ORIGINALITY 221
13.1 Proof of Copyright Ownership Using Digital Time stamp in Malaysia 221
13.1.1 Introduction 221
XV
13.1.2 Related Works: Copyright 223
13.1.3 Related Works: Methods of Proving Copyright Ownership 224
13.1.4 Digital Timestamping Scheme to Prove Copyright Ownership 227
13.1.5 Malaysia Digital Signature Act 1997 and Regulations 1998
(Act 562) 227
13.1.6 Malaysia Evidence Act 1950 (Act 56) 230
13.1.7 Security Analysis 231
13.1.8 Computation Load 232
13.1.9 Implementation Costs 232
13.1.10 Comparisons 234
13.1.11 Conclusion 235
13.2 IP Evidence Using Digital Timestamp 235
13.2.1 Introduction 235
13.2.2 Related Works: Patent Filing Systems 236
13.2.3 Related Works: Digital Timestamping Scheme 237
13.2.4 Proposing Limited First-to-invent Patent Filing System 237
CHAPTER 14: HACK-PROOF DATA STORAGE USING
INNOVATED DIP SWITCH 241
14.1 Abstract 241
14.2 Introduction 241
14.3 Related Prior Arts 242
14.4 Proposed Secure DIP Switch 243
14.5 Method and Device to Secure Hacking-Free Data Storage 245
14.6 Costs and Reliability 246
14.7 Other Forms of Innovation 248
14.8 Conclusion 249
CHAPTER 15: CONCLUSIONS 250
15.1 Concise Summary 250
15.2 Other Supporting Reading Materials in This Research Project 252
15.2.1 NIST Publication 252
15. 2. z Security and Privacy 253
15.2.3 Other Resources 253
15.3 Suggestions for Future Research 254
15.3.1 Fixed- Width Font Supporting All the Unicode Graphic Symbols 254
15.3.2 512-Bit Multihash Key Needs Hash Function beyond 1024 Bits 254
15.3.3 MePKC Extension to Other Non-Conventional Cryptographic
Schemes 255
15.3.4 Big Secret(s) for Information-Hiding and Non-Cryptographic
Applications 255
15.3.5 Safety Box Using Computerized Lock 256
15.3.6 Provable Security Studies 256
15.3.7 Statistical Surveys for Various Security Schemes 257
xvi
15.4 Conclusions 257
APPENDIX A: WRITING SYSTEMS OF THE WORLD 258
APPENDIX B : CHILD-MADE 2D KEYS 261
REFERENCES 262
ACRONYMS 334
xvii
LIST OF TABLES
Table 3.1 Minimum symmetric key sizes for different security levels of protection
18
Table 3.2 Minimum asymmetric key sizes in equivalent with the security levels of
symmetric key sizes 19
Table 3.3 Various key sizes corresponding to the numbers of ASCII characters,
Unicode (version 5.0) characters, and password units of various secret
creation methods, as well as the settings sufficiency of some key input
methods and systems 21
Table 4.1 Passphrase generation from acronym 31
Table 4.2 Minimum diceware words (7776 word list) for different security levels
33
Table 4.3 Conversions between binary and hexadecimal numeral systems 33
Table 4.4 Key size equivalence for symmetric and asymmetric crypto systems (bit)
35
Table 4.5 Minimum coinware words for various word list sizes (WLS) 35
Table 4.6 Environ password 37
Table 4.7 Phonetic encoding of Hanyu Pinyin (Mandarin-based) 39
Table 4.8 Character structure encoding of Sijiao Haoma 40
Table 4.9 Forms of Romanized Chinese-character-encoded words 40
Table 4.10 Minimum coinware words for Han character combined list (70229
words) 41
Table 4.11 Forms of self-created Chinese language passphrase for ( 汉 ) 43
Table 4.12 Phonetic encoding of jyutping in Cantonese language 44
Table 4.13 Forms of self-created Cantonese language passphrase for (汉) 45
Table 4.14 Forms of Japanese language password for (大) ， (漢) ， and (山)
46
Table 4.15 Minimum key lengths for various Chinese key spaces (in Han character)
46.
Table 4.16 Dimensions or 2D key for various symmetric key sizes 60
Table 6.1 Frequency analyses of various sources [ascribe to Bauer (2002)] 94
xvm
Table 6.2 Sequence of keyword rules for modified frequency analysis 97
Table 6.3 ASCII mutual substitution table 105
Table 6.4 Determination of the critical points of function j2 107
Table 7. 1 Binary-to-text encoding Bin2Txt(H) of multihash key 122
Table 7.2 Comparisons of key management tools 124
Table 8.1 Comparisons of classical and multihash-based digital signature schemes
154
Table 13.1 Comparisons of our scheme with linking and tree timestamping schemes
232
Table 13.2 Cost estimation table of a digital timestamping scheme in Malaysia on 8
June 07 233
Table 13.3 Comparisons of methods to prove copyright authorship/ownership 234
Table 13.4 Advantages and disadvantages of first-to-file patent filing system 239
Table 13.5 Advantages and disadvantages of first-to-invent patent filing system 239
Table 13.6 Advantages & disadvantages of limited first-to-invent patent filing
system 240
Table 14.1 Operating modes of method and device to secure hack-proof data storage
246
Table A. 1 Functional classification of writing systems 259
Table A.2 List or languages by number of native speakers 260
xix
LIST OF FIGURES
Figure 0.1 Two-dimensional (2D) key x
Figure 3.1 Generations and applications of one/more big memorizable secrets 28
Figure 4.1 Chinese poem for easy memorization of Sijiao Haoma 40
Figure 4.2 Example of self-created signature-like Han character by modifying the
Han character of (汉) from [Hanyu Pinyin = han4] and [Sijiao Haoma
= 37140] to [Hanyu Pinyin = han4] and [Sijiao Haoma = 37141] 42
Figure 4.3 Example of self-created signature-like Han character from {han} ( 、汉 )
48
Figure 4.4 Operation of 2D key input method and system 52
Figure 4.5 Pseudocode of 2D key input method and system 55
Figure 4.6 Styles of 2D key: Multiline passphrase 57
Figure 4.7 Styles of 2D key: Crossword 57
Figure 4.8 Styles of 2D key: ASCII art 58
Figure 4.9 Styles of 2D key: Unicode art 59
Figure 5.1 One of the exemplary tabular pages of multilingual key consisting of the
first 2^o Han characters in the Unicode and starting from Unicode value
{4E00} 63
Figure 5.2 A Han character from Unicode before and after the grid partitioning for
various settings 64
Figure 5.3 Grid partitioning encoding of a graphic symbol 65
Figure 5.4 (16+l)-colour scheme for colourful multilingual key 68
Figure 5.D Operation of multilingual key input method and system 70
Figure 5.6 Operation of multi-tier geo-image key input method and system 73
Figure 5.7 Software token generation of multi-factor multimedia key input method
and system 75
Figure 5.8 Software token acquisition and application of multi-factor multimedia
key input method and system 76
Figure 6.1 Al-KindT's explanation on the revolutionary system of cryptanalytic
technique 92
Figure 6.2 Implementation of enhanced frequency analysis 98
XX
Figure 7.1 Pseudo-code to determine the numbers of hash iteration for multiple
security levels of multihash key methods and systems 119
Figure 7.2 Operation of the basic model of multihash key method and system 120
Figure 7.3 Proposed usages of 20 security levels 123
Figure 7.4 Methods and systems to support more offline accounts for multihash
130
Figure 7.5 First variant of multihash key method and system to support more offline
accounts using automatically selected tiers and security levels 137
Figure 7.6 Second variant of multihash key method and system to support more
offline accounts using automatically selected permutation sequence of
security levels 139
Figure 7.7 Third variant of multihash key method and system to support more
offline accounts using a hybrid combination of automatically selected
tiers and security levels, and automatically selected permutation sequence
of security levels 142
Figure 7.8 Fourth variant of multihash key method and system for the specific
application to act as a further authentication factor in the Internet banking
or other situations 144
Figure 7.9 Fifth variant of multihash key method and system for the specific
application to act as a simple key escrow method and system for
supervisor-wise non-critical secrets 145
Figure 8.1 Pseudocode of the classical digital signature scheme 149
Figure 8.2 Pseudocode of the proposed multihash signature scheme 151
Figure 8.3 Multihash signature method and system to provide object-designated
signature message 153
Figure 9. 1 Operation of MePKC method and system 161
Figure 10.1 Data embedding process into a cover data for method and system to
harden the identification of an embedded data in steganography although
stego-data has been detected 166
Figure 10.2Data extracting process of embedded data from a stego-data for method
and system to harden the identification of an embedded data in
steganography although stego-data has been detected 167
xxi
Figure 10.3 Samples of digital cheque in triple-watermark digital cheque scheme 169
Figure 10.4 Creation of blank cheque by a bank and written cheque by a payer in the
triple-watermark digital cheque method and system 170
Figure 10.5 Cheque crediting process by a payee in the triple-watermark digital
cheque method and system 171
Figure 10.6 Samples of digital software license in triple-watermark digital software
license scheme 174
Figure 10.7 Creation of blank software license by a vendor and written software
license by a reseller in the triple-watermark digital software license
method and system 175
Figure 10.8 Endorsement process of a software license by a licensee in the triple-
watermark digital software license method and system 176
Figure 11.1 Various not- so-frequent operations of the basic model of MePKC
authentication schemes with feature of non-plaintext equivalence 181
Figure 11. 2 First basic model of MePKC authentication scheme between a human
user and a computer with features of non-plaintext equivalence and
optional mutual authentication 1 82
Figure 11. 3 Various not- so-frequent operations of the second model of MePKC
authentication schemes with features of non-plaintext equivalence and
perfect forward secrecy 183
Figure 11. 4 Second model of MePKC authentication scheme between a human user
and a computer with features of non-plaintext equivalence, perfect
forward secrecy, and optional key exchange scheme (Part 1) 184
Figure 11.5 Second model of MePKC authentication scheme between a human user
and a computer with features of non-plaintext equivalence, perfect
forward secrecy, and optional key exchange scheme (Part 2) 185
Figure 11. 6 MePKC digital certificate with four public keys for various applications,
such as password throttling 188
Figure 11. 7 Three-tier MePKC digital certificates for various applications, such as
persistent private key, rolling private key, and ladder authentication 191
Figure 12.1 Operations to record, store, access, manage, and download the voice
mail, voice call, and video call in the distributed servers at the CO of
PSTN of wireline phone and/or CM of MTSO of wireless phone 196
Figure 12.2 ANN based BAP and its smallest model of 4-node distributed network
201
Figure 12. 3 Total number of exchanged messages for different types of BAP 202
Figure 12.4Partitioning of a distributed network and its optimal partitioning selection
203
Figure 12.5 Partitioning of the entities involved in the electronic commerce
transactions into three groups: Essential group, government group, and
non-essential group 204
Figure 12.6Tripartite ANN based BAP with trusted party and faulty node detection
for multipartite electronic commerce transaction using MePKC
cryptographic schemes for communications 205
Figure 12.7 Tripartite ANN based BAP without trusted party but still with faulty
node detection for multipartite electronic commerce transaction using
MePKC cryptographic schemes for communications 206
Figure 12. 8 Group efficiency of a committee meeting according to the Kurokawa's
human interaction model 211
Figure 12.9 Group efficiency of an exploratory group according to the Kurokawa's
human interaction model 212
Fig. 12.10 Success probability of technology transfer according to the Kurokawa's
human interaction model 213
Fig. 12.11 Group efficiency of an exploratory group formed from leaders of some
committee meetings (without condition for common consensus) as
modified and enhanced from the Kurokawa's human interaction models
214
Fig. 12.12 Group efficiency of an exploratory group formed from leaders of some
committee meetings (with condition for common consensus) as modified
and enhanced from the Kurokawa's human interaction models 215
xxiii
Fig. 12.13 Success probability of an exploratory group formed from leaders of some
committee meetings (with condition for common consensus) ) as
modified and enhanced from the Kurokawa's human interaction models
216
Fig. 12. 14 Method and system to boost up the trust level of MePKC digital
certificate by using more than one certification authority (CA) and/or
introducer of trust of web 219
Figure 13.1 Generation of timestamp and superhash 228
Figure 13.2 Verification of a signed timestamp 229
Figure 13.3 Verification of daily superhash to limit TSA power 229
Figure 14.1 Structural diagram of conventional 10-way DIP switch 244
Figure 14.2 Structural diagram of proposed 10/ 12- way secure DIP switch 244
Figure 14. 3 Innovated 10-way 8PST+2PST DIL switch activated in opposite
direction 248
Figure A.l Writing systems of the world 258
Figure B.l 2D keys using ASCII art and Chinese characters meaning "twenty first
day" [二 ^ ^一 日] 261
Figure B.2 2D keys using ASCII art and Chinese characters meaning "cloudy sky
nurtures the woods" [云 天工 木] 261
xxiv
PREFACE
This book of research essay in the form of thesis towards a future possible
doctorate thesis is the author's works prepared outside the working time of
undergraduate teaching. It is published online here mainly for advertisement
purposes to save development costs, public peer reviews due to the lacking of
evaluation experts in the author's social networks, and to speed up the research
process of human civilization for the betterment of multicultural societies.
In addition, in parallel with the purpose to gain public peer review, this book
will be an output documentation of a research project registered on 27 May 2004 to
achieve three aims at one stroke. These three aims are to solve an imperative research
problem, to develop intellectual properties (IPs) to support an entrepreneurship, and
to qualify a person for a doctoral degree. The proposal defence seminar, first work
completion seminar, second work completion seminar, notice of thesis submission
request, and MMU (Multimedia University) approval of this thesis title to enable its
experts' evaluation are on 14 March 2005， 18 February 2008， 2 July 2008， 23 July
2008， and 1 December 2008， respectively. Nevertheless, this thesis-like book is
mainly prepared in October 2008.
Before reading further the contents of this book (aka thesis), please be
reminded that the process to produce the research results of this thesis is full of
human-made obstacles like graft, office politics, supervisor appointment, placement
of doctoral candidature under closer monitoring on the brink of termination, oral
requests to give up the teaching assistantship in the form of sound snatching to stop
the doctoral candidature, listings of authorship and inventorship, procrastination of
work completion seminar, extraordinary request of another work completion seminar
even though sufficient novel research results have been presented, etc.
Hence, to those whom have significantly and purposefully caused the
damages and delays in the production of any author's research outputs, like a cash
value of MYR$200-00 and procrastination of 1 hour, please confess to the God(s)
and surrender yourself to the judge(s) for deserved punishment before proceeding
with the thesis contents and any applications for any purposes. Otherwise, for the
breach, please damn yourself as well by God(s) for the deserved punishment together
XXV
with your offspring supporting your deeds. And yet for those who are helpful for
their kind-hearted deeds towards the success of this thesis book, may the God(s)
blesses them and their offspring supporting their deeds.
Due to the relatively high research costs invested by the author, for refund, as
well as for building up a fund for further maintenance, research and development,
any original and novel idea conceptions from the author in this book is only free of
usage for public interests, press report, private study, research, and teaching
throughout the World, with the condition that proper originality citation for source
references has been clearly shown. Yet for any commercial usage, prior consent has
to be obtained from the author or his successor(s).
After the author's decision for not furthering his doctoral research studies
under the Lee Foundation Scholarship as communicated by Professor Michael T.-C.
Fang, this research project began with its idea conception in the end of 2003 by
having the official PhD project application date on 12 November 2003. It began with
the studies of multimedia communications security in general and autosophy
communications in particular. Then, in October 2005， some novel ideas were
conceptualized on how to protect the data crystal of autosophy communications in
particular, which is then generalized for any common computer data protection, to
networked information security, and any applications of big secret beyond 128 bits in
information engineering.
Let's create and maintain a networked info-computer age for a more
paperless, petroleum-less, and environment-friendly human society by having safer
multipartite electronic computer communications as from the original and novel
knowledge contribution of this research project.
xxvi
CHAPTER 1 OVERVIEW
1.1 Introduction
The world human population in year 2008 has achieved beyond 6.7 billion.
At the same time, the world climate, resources, and environment are having red
alarms on. Information communications technologies (ICT), especially the electronic
communications of Internet, are believed to be tools to reduce the paper usages and
transportation demands, as well as to cultivate a global economy with smoother
demands and supplies. Security, health, food and beverages, accommodation, family,
career, education, finance, sex, entertainment, sport, etc. are human major concerned
topics. Their importance is in descending order for a majority of people.
Here, when ICT is applied to preserve more Earth resources and to conserve
friendly environment, information security is always a major people concern for
important computer communications. As for the Internet, identity theft is a serious
crime in the electronic commerce and electronic government. Yet another serious
offence is copyright piracy of literary works, software, music, image, and video. Due
to the hacked computer databases of human records, the rights of privacy and
publicity are also hard to be controlled and guarded.
1.2 Motivation
In term of information security, it mainly consists of cryptology, information
hiding, and random number generator (RNG). Cryptology further consists of
cryptography and cryptanalysis. Information hiding further consists of steganography
and digital watermarking. RNG further consists of hardware RNG and software
pseudo-random number generator (PRNG).
To access and control a user identity of an information security system, there
are four types of authentication factors: What you know like secret, what you have
like token, what you are like biometrics (Menezes, Oorschot, & Vanstone, 1996;
Boatwright & Luo， 2007)， and whom you know like introducer (Brainard, Juels,
Rivest, Szydlo, & Yung, 2006)， in the ascending order of implementation costs.
1
These factors can be used individually or mixed. Among them, password the
secret is the most prevailing one for applying the symmetric key cryptography in the
Internet due to the low implementation costs, as well as good hardware and software
compatibilities. However, a secret, especially a long one, is subject to the
forgetfulness or the exposure of a secret written down. The situation becomes worse
when there are lots of accounts to be handled. If a secret is used for multiple accounts,
there exists domino effect of password reuse problem (Ives, Walsh, & Schneider,
2004) . Moreover, the memorizability size of a secret using the current prior art is
limited to 128 bits for a protection period of 30 years. To solve these problems, token
or biometrics with optionally bi-factor using a secret is used.
Nevertheless, token has the weaknesses of poor hardware and software
compatibilities, low portability when number of tokens per user is many, high
implementation costs (installation and maintenance), easy loss, possible dropping
damages, and token cracking (de Koning Gans, Hoepman & Garcia 2008; de Winter,
2008; Garcia, de Koning Gans, Muijrers, van Rossum, Verdult, Schreur & Jacobs,
2008; Wikipedia Contributors, 2008az).
Meanwhile biometrics has the disadvantages of poor hardware and software
compatibilities, domino effect due to limited biometrics to support multiple accounts,
no perfect accurate system due to FAR (False Acceptance Rate) and FRR (False
Rejection Rate), low usability and efficiency due to no universal accessibility and no
permanent availability from physiological and medical factors (Maghiros, Punie,
Delaitre, Lignos, Rodriguez, Ulbrich, Cabrera, Clements, Beslay, & van Bavel,
2005) ， high implementation costs (installation and maintenance), as well as
irreplaceability and irreusability problems of biometrics upon hacking and stealth.
For examples of no universal accessibility of biometrics authentication
systems, there is no support for homozygotic twins, 5% of human are not fingerprint
recognition (Haylock, No date; Maltoni, Maio, Jain & Prabhakar, 2003; Vacca, 2007，
p. 280) supported due to diseases like eczema ("Singaporean Female," 2008) and
arthritis, human undergone surgery changing the facial structure needs re-enrolment
for face recognition, 1.8 aniridia patients out of 100,000 births and patients after laser
indotomy to correct angle-closure caused by glaucoma have no iris and are not iris
2
recognition supported, eyes alignment problem with camera of blind people and
patients of pronounced nystagmus (tremor of the eyes) are poorly iris recognition
supported, wheelchair users have usability problems of camera location and
insufficient height variation, cataract patients after operation may need re-enrolment,
and today DNA methods fail to differentiate monozygotic twins. For example of no
permanent availability, the high biometrics deformation rates of very young and very
old require frequent re-enrolment.
For the fourth authentication factor of "whom you know" like introducer and
referee, even though the authentication burden of the introducee can be relieved, the
burden has in fact been transferred to the introducer and it is up to the introducer to
use the authentication factor of what you know, what you have, and/or what you are.
Furthermore, there exist trust, responsibility, and obligation problems between the
introducer and introducee. The human interaction models (Kurokawa, 1988， 1990，
1991， 1997) are then required to analyze the security probability of this factor.
In view of the limitations of token and biometrics, how good if the
weaknesses of secret like memorizability and entropy size can be improved until the
token and biometrics are not needed for majority applications.
1.3 Research Aims
Here, the first main focus of this research project is for this direction:
Methods and systems to create big and yet memorizable secret(s). From a sufficiently
large and yet memorizable master key, it shall be possible to derive multiple unique
slave keys for multiple offline and online accounts. These slave keys shall be
impossible to be used to derive other slave keys.
For public-key cryptography (PKC), the smallest practically secure private
key size is 160 bits by using the FFC (Finite Field Cryptography) or ECC (Elliptic
Curve Cryptography) (Gehrmann & Naslund, 2005， 2006， 2007; E. Barker, W.
Barker, Burr, Polk, & Smid, 2007a, 2007b). Using the current prior arts like
encrypted private key, split private key (Ganesan, 1996b), and roaming private key
(Baltzley, 2000)， there has been no fully memorizable private key yet. Here, the
second main focus of this research project is to develop fully memorizable private
3
key towards MePKC (Memorizable Public-Key Cryptography), aka MoPKC (Mobile
Public-Key Cryptography). The third research focus is to apply the outputs from the
first and second foci of this research project for various applications in the field of
information engineering.
1.4 Research Methodology
This research project originally contributes novel methods and systems to
create big and yet memorizable secret(s) and then MePKC for various applications in
information engineering. As from Spafford ( 1 993)， there are three types of
techniques to prove a model in a computing dissertation: Analytic method using
formal manipulations, stochastic method using statistical measurements, and building
a prototype for experimental testing.
For the research methodology of this research project, the third method of
building a prototype is used to show that it is possible to create big and yet
memorizable secret by using 2-dimensional (2D) key (Lee, 2006b, 2008i)， and
further for the practical realization of MePKC (Lee & Tan, 2006b; Lee, 2008j). Some
prototypes of the other big secret creation methods have also been built (Lee & Tan
2006a; Lee, 2008k).
To show the security strength and protection period of various security
schemes in this research project, the first method of analytically formal
manipulations has been used. Since the author cum researcher is an electrical
engineer and not educated as a mathematician, the reduction-based security (aka
provable security) approach is only tried on his best effort as time allows. Hence,
those cryptographers from the mathematics field are expected to carry out some
provable security studies on the security schemes proposed here whenever the big
secret creation method(s) is applied, especially for the MePKC and its applications.
1.5 Organisation of the Thesis
Generally, the present invention of this thesis generally relates to computer
communications security. More particularly, the present invention relates to key
4
management of cryptography and information security. Most particularly, the present
invention relates to methods and systems to create big and yet memorizable secrets
that are large enough for the higher levels of security strength of security systems
like AES-256, 256-bit ECC, 256-bit PRNG, and so on, (where AES stands for
Advanced Encryption Standard; ECC stands for Elliptic Curve Cryptography; and
PRNG stands for Pseudo-Random Number Generator), together with their derived
applications in the general field of information engineering and specific field of
information security like memorizable public-key cryptography (MePKC).
Particularly, the present invention broadly provides novel generation methods
and systems of big memorizable secrets to practically realize stronger security levels
of cryptographic, information-hiding, and non-cryptographic applications in the
information engineering, especially MePKC (Memorizable Public-Key
Cryptography). The first independent embodiment of the present invention is the
methods and systems to create big and yet memorizable secrets. The second
independent invention embodiment is mutlihash key using hash iteration and hash
truncation to create multiple slave keys from a single master key. And yet the third
independent embodiment of the invention is multihash signature that allows object-
designated message with specific meaning, function, or recipient. From these three
independent inventions, there are then various types of dependent inventions for
various practical applications mainly due to the existence of big memorizable secrets.
The organisation of this thesis has three components: Preliminary section,
chapter section, and postscript section. The preliminary section consists of front page,
copyright page, declaration, acknowledgements, dedication, abstract, table of
contents, list of tables, list of figures, and prefaces. For the postscript section, it has
an appendix to show the writing systems of the world, references, and acronyms.
There are 15 chapters in the chapter section. Chapter 1 is an overview of this
research project. Chapter 2 discusses the life of research postgraduate students in
brief, general conditions to get a doctoral degree, and the practical evaluation on
knowledge contribution to qualify a person as a quality researcher.
Chapters 3， 4， and 5 form closely related contexts to explain methods and
systems to create big and yet memorizable secret. These methods include Chinese
5
language passphrase (CLPP), 2-dimensional (2D) key, multilingual key, multi-tier
geo-image key, multi-factor multimedia key using software token, and their hybrid
combinations. Chapter 6 enhances the secret randomness by using multimedia noises.
Chapter 7 presents a method and system called multihash key to generate
multiple unique slave keys (aka site keys) from a master key for both offline and
online accounts. This multihash key uses hash iteration and hash truncation. Every
slave key is computationally infeasible to be used to derive another slave key.
Chapter 8 on multihash signature uses the concept of hash iteration again to generate
multiple unique signature of a message file from only a single pair of asymmetric
key. Both multihash key and multihash signature are integrated into various MePKC
applications as explained in the latter chapters.
Chapters 9 to 13 are all on the applications of big secret and MePKC. Chapter
9 includes the applications for symmetric key cryptography, MePKC, as well as
other cryptographic, information-hiding, and non-cryptographic applications.
Chapter 10 includes the identification hardening of embedded data in steganography,
electronic fund transfer, and electronic software licensing. Chapter 11 includes the
local/remote human-computer and human- human authentication without shared
secret, MePKC digital certificate having more than one asymmetric key pair for
different protection periods and password throttling, as well as three-tier MePKC
digital certificate for ladder authentication. Chapter 12 includes archiving the
voice/ video calls of wired/wireless phones, multipartite electronic commerce
transactions, as well as trust boosting of MePKC digital certificate by using more
than one certification authority and/or introducer of trust of web. Chapter 13 talks on
MePKC timestamping scheme for evidence of intellectual property (IP) originality.
Chapter 14 is a hardware contribution to further secure the computer
communications of MePKC from virtual hacking over a connected computer
communications network like Internet. It is about a hack-proof data storage using
innovated DIP (Dual In-Line Package, aka DIL) switch. Chapter 15 concludes this
thesis by giving a concise summary on the originally contributed novel knowledge
and some suggestions for future research.
6
CHAPTER 2 EVALUATION ON KNOWLEDGE
CONTRIBUTION
2.1 Life of Research Postgraduate Students
Due to the abundance of knowledge to be explored since the availability of
computer assistance for the limits outside of the human capabilities, the research
postgraduate students at the universities studying for master's and doctoral degrees
blossom not only in the developed countries but the developing countries like
Malaysia as well. However, many postgraduate students in the developing countries
encounter the suitability problems of closely related research supervision and fresh
supervisors who are lacking of the supervision experience. Hence, it is good for a
postgraduate student to browse some informative materials discussing about the life
of research postgraduate students to smoothen the postgraduate studies.
Callahan (2001) has a brief article telling the story of Ph.D. study in the USA.
Pratt (1997) has a graduate school survival guide for us. desJardins (1994, 1995)
published articles in ACM Crossroads on how to be a good graduate students. Lauer
(No date) talked on how to prepare a Ph.D. thesis proposal in computing science.
Some supervisors require postgraduate students to publish copyrighted
journals and conference proceedings papers as a prerequisite for thesis submission.
Hence, it is good to know how to prepare a first draft of paper (San Francisco Edit,
No date). Smith (1990) told the task of the referee who would review the manuscript.
An IEEE fellow, who is B. K. Bose (2006)， disclosed the tips to publish a
transactions article. Spafford (1993) briefed on Ph.D. dissertation (Wikipedia
Contributors, 2008t). Comer (1993) edited an article on how to write a Ph.D.
dissertation.
2.2 General Conditions to Get a Doctoral Degree
For lengthy and more informative materials on the life of research
postgraduate students, there are some books on this topic (Noble, 1994; Elphinstone
& Schweitzer, 1998; Cryer, 2000; Rugg & Petre, 2004; Phillips & Pugh， 2005).
The first doctoral degree was conferred in Paris circa 1150 (Noble, 1994).
Since then, the most basic condition to get a doctoral degree is to have contributed
sufficiently original and novel knowledge to the society. For honorary doctorate, the
analogous condition is sufficiently significant contribution to the society.
In some universities, the thesis publication is an optional component. Yet in
some universities now, a postgraduate student has to follow some preliminary
courses like entrepreneurship and research methodology. Other possible conditions
are proposal defence of research proposal and seminar defence of thesis. Depending
on research supervisors and universities, publishing journals and conference
proceedings papers are optional conditions to qualify for thesis submission.
2.3 Prior Approaches of Research Evaluation
Just like the quality control and evaluation of products and services, there are
various approaches to evaluate the research output. For doctoral dissertation awards
and novel idea competitions of the universities and professional organizations like
ACM (Association for Computing Machinery), Google, and IEEE (Institute of
Electrical and Electronics Engineers, Inc.), good evaluation techniques of research
results and novel ideas are required to determine who the award winners shall be.
The evaluation of a researcher's contribution is traditionally majority-wise
based on the publication counting, credit of journals and proceedings, citation
counting, H- index (Hirsch, 2005; Batista, Campiteli & Kinouchi, 2006; Wikipedia
Contributors, 2008bg) qualified peer evaluator, and contribution impact. The last
element, which is contribution impact, is in fact the main key measure to evaluate the
contribution of a researcher to the welfare of human civilization. However, this
element is not easy to be carried out by the management without the advice of the
expert. Hence, in deciding the recruitment, promotion, compensation, funding
allocation, reviewer list, and consultancy partnership, the publication counting
method is always used due to its easiness.
Lately, Parnas (2007) published his view on the publication counting method
and called for the halt of number game. He referred a paper by Ren and Taylor
(2007)， which calls for automatic and versatile publication ranking, and pointed out
8
the weaknesses of publication counting method. These weaknesses are encouraging
superficial research, overly large groups, repetition, small and insignificant studies,
as well as rewarding publication of half-baked ideas. Parnas also acknowledged that
publication counting method could corrupt the researchers because malicious
researchers would have the behaviour of publishing pacts, clique building, anything
goes, bespoke research, and minimum publishable increment (MPI). As a solution,
he called for an accurate researcher evaluation, where the researcher's papers were
studied by qualified peer evaluators. The factors of slower time and e valuators'
compensation are the disadvantages of this method.
2.4 Contribution Impact
Here, a refinement of researcher evaluation from the Parnas' proposal is
proposed. Heavier weight shall be given to the contribution impact, which is to be
studied by the qualified peer evaluators. The evaluation methods based on
publication counting, credit of journals and proceedings, and citation counting are
inaccurate. For example, the same research result can be published two times as both
the patent and journal (Rivest, Shamir & Adleman, 1978, 1983). A second case is
that a same idea is published three times as proceedings paper, patent and journal
(MacKenzie & Reiter, 2001c, 2002， 2004). The example of the second case has
another closely related conference proceedings paper (MacKenzie, Oprea & Reiter,
2003). The third case is that an idea is published four times as technical report,
proceedings paper, patent, and journal (MacKenzie & Reiter, 2001a, 2001b, 2003，
2006). One more possible case is that an idea may be published as many as eight
times in the forms of technical report, conference proceedings paper, patent, journal,
thesis, book chapter, and book.
The proposal here is mainly aimed at evaluating the research results of the
scientists and engineers. The research types are fundamental, applied, and design-end
engineering. Meanwhile, the research results can be discovery, innovation or
invention documented as one or more types of IPs (Intellectual Properties). It can be
copyright like technical report, proceedings paper, magazine, letter, journal, book,
thesis, computer program, architectural plan, etc. Conference proceedings papers
9
include articles in the conference, symposium, workshop, and colloquium. Other
forms of IP are patent, utility model, industrial design (aka design patent), layout-
design of integrated circuit, trade mark, and confidential information. Two more
intangible assets are publicity right and privacy right.
Proceedings paper, magazine, letter, journal, and thesis are normally
reviewed lightly within a short time by one to three peer evaluators. On the other
hand, patent, utility model, and industrial design are evaluated intensively by patent
examiner within a long period of time ranging from 1 to 6 years or more. To fulfil
the patentability, there are four conditions: Novelty, utility (aka industrial
applicability), non-obviousness, and specification (i.e. within the patent classification
of a country). Some research results like evaluation, survey, analysis, case study, and
commentaries are not patentable. Commentaries include on-site experience,
viewpoint, technical opinion, and forum.
Besides, the guidelines for inventorship and authorship (Devenport, 2005;
Albert Einstein College of Medicine of Yeshiva University, 2008; Wikipedia
Contributors, 2008ao) shall be enforced strictly. The inventorship is enacted where
only people with idea conception can be listed and not the people with reduction to
practice, which when listed can acquit the accused infringers from infringement
charges (Setty & Gentry, 2002). For authorship, the law is not yet enacted and there
exist only ethical rules. Hence, there are authors due to idea conception, analytic
evaluation, reduction to practice, novel supervisory advice, nominal supervision,
fund raising, etc., where some of them are honorary authorship not accepted by some
editorial offices (IEEE Publications, 2002).
For contribution impact, the successful implementation and realization of a
research result for public usage and application are very important. Many research
results are only documented without a further development. This is in fact no
positive impact at all but negative impact due to a waste of resources.
Hence, besides Parnas' proposal to request a few peer reviewers to evaluate a
researcher's selected papers, the evaluated researcher has to list out the current
impact of one's research results. For paper with multiple authors or inventors, the
specific researcher has to point out one's contribution in the listed paper for impact
10
evaluation. For example, article like this (Re, Borean, Bozzi, et al.， 2002) requires
detailed elaboration on the particular contribution of every single listed inventor or
author. This is important to avoid the pseudo-amplification phenomenon where a co-
written article is considered as whole unit in the invention counting of an inventor or
publication counting of an author.
The contribution impact can be a core, supporting element, and influence to a
past, present, and future product and/or service useful to the human society. The
obsolete technology, replacing technology, and alternative competing technology
have also to be accompanied with the researcher's selected papers.
Modern society has evolved from agricultural and industrial economy into
knowledge economy. In 2004, Taplin (2004) reported in a paper that IP assets
accounted for up to 70% and 40% of market values of all corporate assets in the USA
and Japan, respectively. Chandran (2007) stated that for any company, and especially
for pharmaceutical companies, IP is more valuable than any of its tangible physical
assets, where IP constituted more than 80% of the total revenues of any company
while tangible assets accounted for only 20%. Again, Ocean Tomo LLC (Ocean
Tomo, No date; Wikipedia Contributors, 2008d) figured out that the components of
S&P 500 market value in 2005 also had 79.7% to be intangible assets. Among the IP,
patent has the largest economic value.
From Abril and Plant (2007)， only 5% of the US patents were licensed and
only 3% generated royalties. Research results with commercial values are normally
filed as patent today for refunding research costs, usage control, enabling mass
production, affordable cost and price control, reasonable income, technology control,
official government records, archival for future generations to further study, etc.
Under the new rules of USPTO (Dudas, 2007)， a patent can only carry a
maximum of 5 independent claims and maximum 20 dependent claims in total. For
every independent claim, if fully developed, explained, and evaluated, it can amount
to a proceedings paper and/or journal. Furthermore, inventorship is stricter than
authorship, only the researcher with novel conception can be listed and not the
researcher carrying out the reduction to practice. Hence, the inventorship can better
determine the main original knowledge contributor.
11
Therefore, the Parnas' suggestion has to include the request for the evaluated
researcher to list out the impacts of one's research papers to the human civilization.
Listed impacts by the researcher can speed up the evaluation process. This additional
step can also encourage more research results to be fully developed for public usage,
instead of just discussing a war on some plain papers.
2.5 Research Quality
When the researcher has listed out the implemented products and services as
the impacts of one's research papers, the research quality of these outputs started as
an idea is then to be evaluated. The idea evaluation criteria based on Google Project
10 100 (2008) are reach, depth, attainability, efficiency, and longevity. These five
criteria correspond with five questions "How many people will this idea affect?",
"How deeply are people impacted? How urgent is the need?", "Can this idea be
implemented within a year or two?", "How simple and cost-effective is your idea?",
and "How long will the idea's impact last?", respectively.
Yet there are two persons telling how important a good idea is. Pierer studied
law and economics, and he is the president and CEO of Siemens AG. Oetinger
studied political science in Berlin and business administration at Stanford Graduate
School of Business, and he works for Boston Consulting Group, where he is a senior
vice president and director of The Strategy Institute. In their book entitled "A
Passion for Ideas: How Innovators Create the New and Shape Our World" (Pierer &
Oetinger, 2002)， they wrote in the book cover that "The creation, implementation,
and sustainability of new ideas is the lifeblood ensuring the growth and viability of
any organization. Without continuing innovation, competitive advantage and global
market share are endangered. Once-thriving organizations can find themselves
unprepared for the future." This tells how important new ideas are to any
organization including company, university and government for their thriving
sustainability and continuous success. In a smaller scale, the successful completion
of a research degree like master's and doctoral degrees also depends on the sufficient
novel ideas contributed to the pool of human knowledge for further betterment of our
civilization.
12
In the more developed countries, a good idea resulted from the research is
normally filed for patent before any public disclosure or within one year from the
first date of the public disclosure. Good examples can be observed from the fields of
fibre optic communications (Hayes, 2001, pp. 1-12), laser (Taylor, 2002)， and
cryptography (Rivest, Shamir & Adleman, 1983). For less developed countries, the
IPs like patent and copyright are not so strongly focused and emphasized as in more
developed countries. The reasons are explained by Scalise (1999) that less developed
countries respond to their comparative advantage in imitation by lowering IPR
protection to reduce the cost of imitation's inputs, and more developed countries
respond to their comparative advantage in innovation by raising IPR protection to
increase the value or innovation's output.
Hence, to catch up with the more developed countries, the less developed
countries have to culminate in their research to produce the most dominant form of
IP, i.e. patent. Turk (2005) expressed that the most important IP legal issues of the
twenty- first century is to standardize global IP rights to facilitate a better mass
production, lower operating costs, and smoother supply-demand chains.
The patent grant rate in the US is about 75% (Ebert, 2004， 2005， 2007). The
high patent grant rate does not mean there are lots of useful and yet competitive
inventions. In fact, Professor Adam Jaffe in his testimony before the Judiciary
Subcommittee on the Courts, the Internet, and Intellectual Property claimed that lots
of US patents were worthless and unimportant ("Prepared Testimony," 2007). This
fact is also true for copyrighted publications or journals and conference proceedings.
Then, this comes to the question where the great inventors have gone (Bessen,
2004). However, fame is just one of the main factors people filing patents. Another
more important factor is the economic values of a patent to perform like a property
(Meurer & Bessen, 2008). This has caused an abundance of patent applications in the
US and examination becomes a problem due to the lacking of patent examiners
("Prepared Testimony," 2007). The abuse of US patent system is called to a halt
(Lemley & Moore, 2004; Lemley & Sampat, 2007). To solve the examination
problem, Beth Noveck created a voluntary system called "Peer to Patent Project" for
community patent review to assist the USPTO patent examiners (Duane, 2008; Oram,
13
2008; "Peer to Patent," No date). Later, the JPO (Japan Patent Office) has also
launched its version of community patent review system called
Komyunitipatentorebyu [コ ミュ ニ ティ パ テン ト レビュー]. Lately, the US
Chamber of Commerce (USCOC) has recommended the Peer to Patent Project for
further implementation ("Recommendations for consideration," No date). In analogy,
the journals and conference proceedings also have a similar review system called
Computing Review under the ACM ("Computing Reviews," No date).
The inventions in this research project in the IP form are mainly software
patents or CIIs (Computer- Implemented Inventions). In the US, there has been a time
where people questioning on whether software patents shall exist (Samuelson, 1990).
Now, the question has changed to how to improve software patent quality (Thatcher
& Pingrv, 2007). Here, the inventions of this research project have been filed for
patents (Lee, 2008h) and now are pending for patent examinations. The
implementation for public usages has been partially done. However, the research
quality is yet to be investigated.
To evaluate the research quality, a currently common method is citation
counting. This is true for journals, conference proceedings as well as patents. There
are also citation analyses or journals, conference proceedings, and standards based on
the citing patents. Example is IEEE papers based on US patents (Thomas &
Breitzman, 2006a; Piatt, 2006a, 2006b; Breitzman, 2007， 2008) and EU patents
(Breitzman, 2006). From these IEEE papers, Canning (2006) suggested the libraries
to optimize the financial budgets for scholarly collections. IEEE also has occasional
published articles to analyze the US patents (Goldstein, 2006; Sweet, 2007).
Besides, there are some organizations specialized in the research evaluation
("1790 Analytics," No date; "Research Evaluation," No date). The 1790 Analytics
discovered that hot patents were normally linked with government-funded scientific
research (Thomas & Breitzman, 2006b). Hot patents are patents whose impact on
recent technology developments is particularly strong. These hot patents shall be
identified to know the most competing patents for commercial profits. The damage
award (Aharonian, No date) of a patent infringement has a record of USD$1.52
billion for Alcatel-Lucent v. Microsoft in the US (Wikipedia Contributors, 2008an).
14
Spours (2006) discussed on how to exploit patents for profit. Since the
academicians' research funding at the universities is normally financed by the
taxpayers' money, it shall not be wasted without any refund or profit. To encourage
the commercialization of university research, the US Bayh-Dole Act (Bayh, 2006)
was enacted in 1980 to allow the researchers and universities to be the IP assignees.
Bar ham and Foltz (No date) investigated the patent activities and commercialization
efforts of university life science researchers.
Bray and Lee (2000) analyzed the university revenues from technology
transfer and concluded that the average equity from a university spin-off company
was more than 10 times the average annual income from a traditional license.
Another available data of university licensing revenue is from School of Medicine,
Johns Hopkins University, Baltimore, Maryland, USA ("Licensing Revenue," No
date). From this encouraging university patent revenue, it also indirectly tells that the
research output is practical, useful, and cost-effective for the industries to apply.
Branstetter and Ogura (2005) had an article to show the example of knowledge
spillover from the bio-science academia to the industry.
For the specific patent quality, Malackowski and Barney (2008) had a general
discussion on this topic. On the other hand, Ocean Tomo (No, date; Wikipedia
Contributors, 2008d) has an Ocean Tomo 300 Patent Index to show that patent
quality can be measured.
There are also some patents on how to analyze and select good patents due to
their practical usability and competitiveness (Breitzman & Narin, 2001; J. A. Barney
& J. R. Barney, 2003; Williams, 2004). Patent exists on how to track and audit IP of
open source software (Donner, 1999， 2001; Bonnet, Baroniunas & Webbink, 2008).
There is another patent telling a method to obtain investment income based
on the capitalization of intellectual property (Elliott, 2007). Yet, there are patents on
how to license intellectual property assets (Shelton, 2007)， how to do intangible
property transaction and leaseback business method (Walker, 2007)， and protection
against the changes of intellectual property value (Risen & Covello, 2000).
All of these hint that optimum supply-demand chains or intangible assets are
forming gradually from problem, research, solution, investment, patenting,
15
capitalization, licensing, and commercialization to the customers. It also tells that the
method to evaluate and apply the novel idea, especially patent, is getting better in
terms of relevancy, accuracy, systematization, and valuation. The inventions in this
research project are expected to undergo these processes in the coming decades.
As of today, the best research evaluation experts are from 1790 Analytics
LLC (No date), USA, doing the patent analysis and intellectual property evaluation.
Its services are recruited by the IEEE, USA, which is the top organization in the field
of electrical and electronic engineering, as well as financial experts for equity
investment and other top ranking organizations.
Lastly, a reminder here is to note and emphasize that under the IP laws, only
those persons who have created sufficiently novel knowledge contribution are
qualified to be coined as the originating IP contributors like inventor, designer,
author, etc. This point is important to get rid of the malicious supervisors whom have
no significantly novel knowledge contribution but wanted to be listed as co-
contributors to seek for better career profile towards more bonus and salary
increment, faster and easier promotion, as well as higher social recognition for
awards and reputation.
16
CHAPTER 3 CREATING BIG MEMORIZABLE SECRET
(PART 1)
3.1 Required Protection Periods and Their Key Sizes
According to Kerckhoffs' law (Schneier, 1996)， a crypto system shall depend
100% on the secrecy of password or key only. In the words of Shannon's maxim, it
means "enemy knows the system". This law makes the civilian crypto system to have
publicly known algorithm except the classified governmental and military
information. This is needed to gain the public confidence for general daily
applications from the fear of possible backdoor. There are various applications of
secret in information engineering. Here, the required protection periods and their key
sizes are briefly discussed to know how big a memorizable secret shall be.
If a cryptographic algorithm is securely tested, the required key length in
character (L c ) of a password will depend on the factors of number of characters (C)，
key space (S)， secure period (T)， guesses per unit of time (G)， and probability of
guessing (P) (US Department of Defense, 1985). The minimum key length has to be
able to resist the brute force attack. The relationships of L c ， C， S， T， G and P are
given in Equations (3.1) and (3.2).
^=f ( 3 .i)
L「 >
log 2 C
(3.2)
Nowadays, character encoding of ASCII is the most popular computing code.
ASCII has some key sets of 26 lowercase characters, 26 uppercase characters, 10
digits, 62 alphanumeric characters, 33 non-alphanumeric characters, 95 printable
characters, etc. If a key only has symbols of digits, its specialized name is passcode.
If a key is long or consists of printable characters, it is named as passphrase.
There were once three Data Encryption Standard (DES) challenges as in year
1997， 1998， and 1999. Using the distributed network computing, maximum guesses
of 2.45xl0 9 keys per second was once recorded. For the latest guesses per computer
17
as at end of year 2005, it is about 1.5x10 keys per second. The increment rate
follows the Moore's Law (Wikipedia Contributors, 2008v) where computer
performance is doubled for every 18 months. This indicates that strong password has
to be longer as time passing by if there is no special key processing added.
Key length in bit (L) means that there are 2 n possible keys for n-bit key. By
year 2010， the required key is 80 bits for symmetric key algorithm as announced by
U.S. National Institute for Standards and Technology (NIST). Meanwhile,
asymmetric key algorithm, like RSA, needs 1024 bits to be equivalently strong with
80-bit symmetric key algorithm as claimed by RSA Security. The key space varies
and depends on the security requirements.
For the AES suggested by NIST to replace the DES, it has three types of
symmetric key sizes. These key sizes are 128， 192, and 256 bits. Therefore, we have
AES-128, AES-192, and AES-256 to fulfill the demands of security levels at 128，
192， and 256 bits (E. Barker, W. Barker, Burr, Polk, & Smid, 2007a, 2007b). For
other security levels at 80 and 112 bits, NIST suggested two -key Triple Data
Encryption Algorithm (2TDEA) and three-key Triple Data Encryption Algorithm
(3TDEA), respectively (E. Barker, W. Barker, Burr, Polk, & Smid, 2007a, 2007b).
Table 3.1 Minimum symmetric key sizes for different security levels of protection
Key Size, bit
Protection
# 1
32
Individual attacks in "real-time". Only acceptable for authentication tag size.
#2
64
Very short term protection. Obsolete for confidentiality in new systems.
#3
72
Short to medium term of protection depending on organization size.
#4
80
Smallest general purpose level, < 5 years protection.
#5
112
Medium term protection. About 20 years.
#6
128
Long term protection. Good, generic application independent
recommendation, about 30 years.
#7
256
Foreseeable future. Good protection against quantum computers.
Password choice depends on the strength and memorizability. Strength
depends on key size in bit. Memorizability depends on number of memorized secrets
in a human brain. For minimum key sizes at different security levels, it is shown in
Table 3.1 (Gehrmann & Naslund, 2005， 2006， 2007).
18
For short term memory of English-based digit, Miller (1956) showed an
average of 7 items plus or minus 2 (7 ± 2) (Jones, 2002). The good option is longer
key size in bit and still memorizable. Some articles on memory can be referred
(Baddeley, Thomson & Buchanan, 1975; Ellis & Hennelly, 1980; Hoosain & Salili,
1988; Cowan, 2001; Wikipedia Contributors, 20081， 2008as， 2008aw) and we can see
that a user has 6.5 unique passwords in average (Florencio & Herley, 2007)， or 4 to 5
unrelated keys (Adams & Sasse, 1999). These are textual secret; whereas graphical
secret has higher memorizability (Standing, Conezio & Haber, 1970; Standing, 1973).
On the other hand, there are 3 conventional mathematical hard problems used
in asymmetric key crypto system, which is also called public-key crypto system.
These problems are integer factorization problem, discrete logarithm pro Diem, and
elliptic curve discrete logarithm problem. NIST categorizes the applications of these
problems for public-key cryptography as integer factorization cryptography (IFC),
finite field cryptography (FFC), and elliptic curve cryptography (ECC), respectively.
IFC has a long key size for public and private keys. FFC has a long public
key and a short private key. ECC has a short key size for public key and private key.
The minimum asymmetric key sizes for IFC, FFC, and ECC in equivalent with the
security levels of symmetric key sizes are shown in Table 3.2 (E. Barker, W. Barker,
Burr, Polk, & Smid， 2007a, 2007b).
Table 3.z Minimum asymmetric Key sizes in equivalent with the security levels of
symmetric key sizes
Security
(bits)
IFC
FFC
ECC
Public
Private
Public
Private
Public
Private
80
1024
1024
1024
160
160
160
112
2048
2048
2048
224
224
224
128
3072
3072
3072
256
256
256
192
7680
7680
7680
384
384
384
256
15360
15360
15360
512
512
512
A good password has to be strong and memorizable (Gehringer, 2002). The
random password with printable ASCII characters is the strongest password but it is
19
poor in memorizability (Yan, Blackwell, Anderson & Grant, 2004). However,
password with good memorizability tends to be weak password and under the
password cracking threats of guessing and dictionary attack (Klein, 1990). As time
lapses, longer key length is needed due to the advancement of computer technology.
Hence, the trend is the strong and memorizable passphrase or special key processing
technique like key strengthening is adopted to get rid of the quest of longer key size.
The most popular email encryption software called Pretty Good Privacy (PGP)
9.0 (PGP Corporation, 2006) allows a maximum of 255 characters to be the
passphrase. Microsoft Windows operating systems also have this feature. Methods
exist on how to create secure keys (Adams, Sasse & Lunt, 1997; Brown, Bracken,
Zoccoli & Douglas, 2004). Therefore, it is a research problem here questioning on
how to have big enough and yet memorizable secret(s) for various applications in
information engineering, generally, and security engineering, particularly.
3.2 Review of the Secret for Symmetric Key Cryptosystem
In civilian information security, according to Kerckhoff s Law, a security
system shall depend fully on the secrecy of a key, and not the algorithmic software
nor its hardware. The main reason for this law is that public confidence has to be
earned to show that there is no backdoor in the security system relying solely on the
secrecy of key, and disclosing its algorithm and hardware to the public, especially
academic and corporate researchers, for comments.
For authentication to a security system, it basically has four methods: Secret
for what you know, token for what you have, biometrics for what you own, and
person for whom you know. Due to the factors of cost, hardware and software
compatibilities, password/key the secret is the most popular. Short key is called
password and long key is called passphrase. The key selection is always the balance
of the factors of memorizability and security. Long and random key is securer but
harder to remember. The current prior art of single-line key input field limits the
practical memorizable key size to a maximum of 128 bits for majority normal users.
To create longer password called passphrase, there are now four existing
methods: Sentence-type passphrase, acronym-type passphrase, diceware, and
20
coinware. Sentence-type passphrase is memorizable and has long key size, but
vulnerable to dictionary attack; whereas acronym-type passphrase taking the first,
last, other locations, or hybrid location is memorizable and resists to dictionary
attack, but has a small key size. Diceware and coinware use several dices and coins,
respectively, to randomly select a word from monolingual, bilingual, or multilingual
wordlists, where they can resist dictionary attack, but memorizablity reduces as the
key size becomes longer. Hence, these passphrase generation methods are still
insufficient to create random, memorizable, and yet big secret, that can resist
guessing attack and dictionary attack, to fulfil the need for secret bigger than 128 bits.
Table 3.3 Various key sizes corresponding to the numbers of ASCII characters,
Unicode (version 5.0) characters, and password units of various secret creation
methods, as well as the settings sufficiency of some key input methods and systems
Key size (bit)
80
96
112
128
160
192
256
384
512
Number of ASCII character (6.57 bits)
13
15
18
20
25
30
39
59
78
Number of Unicode character (16.59 bits)
6
7
8
10
12
16
24
31
Number of CLPW unit (85.41 bits)
1
2
2
2
2
3
3
5
6
ASCII-based (4 * 5) 2D key (131.4 bits)
Yes
Yes
Yes
Yes
No
No
No
No
No
ASCII-based (5 * 6) 2D key (197.1 bits)
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
ASCII-based (7 * 6) 2D key (275.9 bits)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Unicode-based (5 * 5) 2D key (414.8 bits)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
BW multilingual key (16.59 bits)
5
6
7
8
10
12
16
24
31
BW multilingual key + grid (19.79 bits)
5
5
6
7
9
10
13
20
26
Color multilingual key (24.59 bits)
4
4
5
6
7
8
11
16
21
Color multilingual key + grid (27.79 bits)
3
4
5
5
6
7
10
14
19
Multi-tier geo-image key (64.82 bits)
2
2
2
2
3
3
4
6
8
Multi-factor key using software token
Halving the memorizable key sizes of MePKC and other
applications at equivalent security levels by using AES.
N.B.: BW = Black-and-white
According to Bruce Schneier (2006, 2007), for a survey of 34,000 MySpace
users' passwords, about 99% of the passwords have 12 ASCII characters or less. An
21
ASCII character carries about 6.57 bits, which means 99% of the 34,000 MySpace
passwords have 78.84 bits or less. This reflects the facts that almost all the
symmetric keys of the current symmetric key crypto systems in practice reach at a
key size less than 128 bits. In other words, memorizable key the secret is only
practically applicable to the current popular symmetric key crypto systems like 112-
bit 3TDES (3-Key Triple Data Encryption Standard) and 128-bit AES (Advanced
Encryption Standard). However, in a large-scale password habit survey (Florencio &
Herley, 2007)， the average password entropy is about 40.54 bits and a user has 6.5
passwords for 25 accounts where 8 accounts in average are used daily.
Table 3.3 shows the numbers of ASCII and Unicode (version 5.0) characters
for various key sizes. In Unicode 5.0， there are 98884 graphic symbols or 16.59 bits
per graphic symbol. The repertoire of Unicode graphic symbols can be upgraded
from time to time in future versions to enlarge the number of graphic symbols.
Memorizable keys for 192-bit and 2^b-bit AES are out of the reach of the current key
management method and system. Hence, need exists to have better key management
method and system to create larger key/password the secret larger than 128 bits.
3.2.1 Related Work: Single-Line Key/Password Field
Conventionally, when secret is used for authentication, single-line key field
will be the area for a user to enter a key. For the current longest possible key, it is a
single-line passphrase. Passphrase can be formed from acronym, sentence, diceware,
and coinware (Lee & Ewe, 2006). Nevertheless, limit exists due to the problems of
memorizability and ASCII character input from keyboard. The first problem is due to
the human factor; whereas the second is due to the user interface. These problems
prohibit the applications of symmetric key sizes at higher security levels, whenever a
user cannot remember and/or conveniently enter a long single-line passphrase.
3.3 Review of the Secret for Asymmetric Key Cryptosystem
Besides the symmetric key cryptography, asymmetric key cryptography or
public-key cryptography (PKC) is one of the two main components in the field of
22
cryptography. PKC emerges in the 1970s (Diffie & Hellman, 1976; Goldwasser,
1997). Symmetric key crypto system has a shared secret key between a pair of users,
but each PKC user has an asymmetric key pair consisting of a private key known
only to the user and a public key shared with the other users. Amazingly, PKC can
solve the key sharing and distribution problems of symmetric key crypto system.
Moreover, PKC can resist the guessing attack, dictionary attack, and pre-computation
attack that symmetric key crypto system is susceptible to. Nevertheless, PKC
processing speea is about 1000 times slower than the symmetric key cryptography.
Consequently, PKC and symmetric key crypto system have to be used in hyond mode
for maximum performance of effectiveness.
Now, there are three main conventional asymmetric cryptosystems: IFC
(Integer Factorization Cryptography), FFC (Finite Field Cryptography), and ECC
(Elliptic Curve Cryptography). IFC is based on the mathematical hard problem of
integer factorization. FFC is based on discrete logarithm problem. And ECC is based
on elliptic curve discrete logarithm problem.
RSA (Rivest-Shamir-Adleman) crypto system is a type of IFC being the very
first practical realization of PKC since 1977. FFC like ElGamal encryption and DSA
(Digital Signature Algorithm), as well as ECC are firstly introduced in the 1980s.
Then, there are other PKC based on different mathematical hard problems but not yet
well- standardized. Nevertheless, so far all the key sizes of asymmetric private key
for IFC, FFC and ECC are too big to be human- memorizable. The large key sizes of
RSA crypto system for its both private and public keys, as well as FFC crypto system
for its public key, have even caused the USA government to shift to ECC having
significantly smaller public and private key sizes. For more details on their
practically secure key sizes, please refer to two NIST articles (E. Barker, W. Barker,
Burr, Polk, & Smid， 2007a, 2007b).
Due to the reason that private key is not fully human- memorizable using the
current prior art, a private key is either fully or partially in the form of a token. In the
mean time among the prior art, there are three basic methods for private key storage:
(i) Encrypted private key stored in the local computing system or device; (ii) split
private key firstly proposed by Ravi Ganesan (1996b) on 18 July 1994 in the US
23
Patent US5557678; and (iii) roaming private key firstly proposed by Cliff A.
Baltzley (2000) on 25 November 1998 in the US Patent US6154543. All the three
methods are bi-factor or multi-factor authentication, where at least one factor is a
secret and another factor is a software token or hardware token.
The first method of private key storage encrypts the private key using a
symmetric key and stores the ciphertext of private key in the local computing system
like hard disk drive or a device like smartcard, floppy disk, or USB flash drive.
Encrypted private key method suffers from the problems of loss, damage, side-
channel attacks, mobility, hardware and software compatibility, and password
domino cracking effect of its digital certificate carrying only one asymmetric public
key.
The second method splits a private key into two or more portions, where the
first portion is a memorizable password or derivable from the memorizable password
kept by the owner of that private key. The second and possible other portions of the
private key are kept by one or more servers in the encrypted form like the first
method. The first, second and possible other split portions of the private key may
also be derived from various authentication factors like token and biometrics. Split
private key method suffers from the problems of malicious central authority attack
on the user's short password, dictionary attack on the stolen encrypted partial private
key, and password domino cracking effect of its digital certificate carrying only one
asymmetric public key.
For the third method, roaming private key also has encrypted private key but
its ciphertext is stored in a network system like server, and owner of the private key
can download it from anywhere and anytime as long as the user has network access.
The roaming private key method suffers from the problems of side-channel attacks,
hardware and software compatibility, malicious central authority, dictionary attack
on the stolen encrypted private key, and password domino cracking effect of its
digital certificate carrying only one asymmetric public key.
In the US Patent US7 113594, Boneh and Franklin (2006) described a new
type of PKC called identity-based cryptography (IDC). In this method, a user's
unique public identity like email or phone number is the public key and hence
24
memorizable. However, its private key is not memorizable and has to be generated
by a trusted third party (TTP).
Notwithstanding , as compared with symmetric key crypto system using
password or key the secret, the popularity of token-based PKC using fully or
partially encrypted private key, is low due to the problems of mobility convenience,
implementation costs, hardware and software compatibilities, and management
difficulty of certificate revocation list. Hence, there exists a need to get rid of fully or
partially encrypted private key, and to invent key input method to let the private key
fully human- memorizable as like the symmetric key.
3.4 Potential Methods to Create Big and Yet Memorizable Secret
One of the many invented methods here to create big and yet memorizable
secret is to innovate the graphical password or picture password. From psychological
studies, it claims that human graphical memory is stronger than human textual
memory. The graphical password is categorized into recognition-based and recall-
based methods by Xiaoyuan Suo, Ying Zhu, and G. Scott Owen (2005). For
recognition-based method, it can be the types of cognometrics and locimetrics.
Meanwhile for recalled-based method, it can be the type of drawmetrics.
Passfaces invented by J. H. E. Davies (1997)， as in the US Patent US5608387,
is a type of cognometircs, where a user is requested to recognize some pre- selected
image sequence of human faces as password. Davies' s method has the weakness of
low entropy per image. For G. Blonder' s method (1996)， as in the US Patent
US5559961, it is a type of locimetrics, where a user has to select a few areas of an
image in sequence as password. Blonder' s method is vulnerable to hot-spot attack
and shoulder- surfing attack. For Draw-a-Secret scheme by L Jermyn, A. Mayer, F.
Monro se, M. Reiter, and A. Rubin (1999)， it is a type of drawmetrics, where a user
draw lines and points on a grid in the form as like a hidden hand signature. For this
Draw-a-Secret scheme, its weakness is its authentication process for either
acceptance or rejection is not exact as in the previous two graphical password
methods, but estimation having FAR (False Acceptance Rate) and FRR (False
Rejection Rate).
25
Besides these three main groups of graphical password, there are icon-like
graphical password scheme by P. V. Haperen (1997)， as in the UK Patent
Application GB23 1 3460, and event-based graphical password scheme by J.
Schneider (2004)， as in the US Patent Application US2004/0250138. The both of
these latter methods are cognometric. Their common weakness is that the key space
or password space is limited by the fine differentiation capability of human visual
memory over images that may have only minor differences. This causes the entropy
per image selection to be still unsatisfactory not big enough for the demands of
information engineering for the stronger security levels to carry more bits of strength.
Hence, there exists a need to boost the key space of graphical password for higher
entropy per image selection and yet still human- memorizable and visually
differentiable.
Another potential method to have big memorizable secret is to create Chinese
language password (CLPW) through Chinese character encodings and their
Romanization. T. D. Huang, as in the US Patent US4500872, proposed on 19
February 1985 to use phonetic encoding and symbolic encoding to represent a
Chinese character. The character space of Chinese language is huge by more than 16
bits per character and yet human- memorizable and differentiable. This CLPW
method can also be extended to other CJKV languages due to the common sharing
for the usages of Han characters (漢字 or 汉字) like Chinese Hanzi, Japanese Kanji,
Korean Hanja, and Vietnamese Han Tu. However, the current CLPW has a weakness
that it is subject to dictionary attack. Hence, there exists a need to create CLPW
resisting the dictionary attack.
There are some inventions to create password that can resist the dictionary
attacks. Among them are (i) "System and Method for Generating Unique Passwords"
by Martin Abadi, Krishna Bharat, and Johannes Marais (2000) in the US Patent
US6141760; (ii) "Password Generation Method and System" by M. R. McCulligh
(2003) in the US Patent US6643784; (iii) "Method and System for Automated
Password Generation" by P. M. Goal and S. J. Kriese (2004) in the US Patent
Application US2004/0 168068; (iv) "Method and Apparatus for Password
Generation" by M. R. Dharmarajan (2005) in the US Patent Application
US2005/0 132203; and (v) "Method and System for Generating Passwords" by B. E.
26
Moseley (2006) in the US Patent Application US2006/0026439. Nevertheless, even
though these five methods can resist dictionary attacks, they have lower
memorizability. Hence, there exists a need not only to have a password generation
method that can resist dictionary attack, but can have high memorizability as well
even for a big secret at least and beyond 128 bits.
Yet another method to create a memorizable secret bigger than the current
prior art was proposed by Whitfield Diffie and William A. Woods (2006) in their
patent application filed on 22 June 2006 entitled "Method for Generating Mnemonic
Random Passcodes", US Patent Application US2007/0300076. However, the
password created by this method is not yet big enough for many applications in the
information engineering.
3.5 Methods and Systems to Create Big Memorizable Secret
Accordingly, the present invention mainly provides some methods and
systems to create big memorizable secrets. These methods and systems include (i)
self-created signature-like Han character; (ii) two-dimensional key (2D key); (iii)
multilingual key; (iv) multi-tier geo-image key; and (v) multi-factor key using
software token. Every method and system can be used individually or mixed as a
hybrid combination. The size of big memorizable secret is at least 128 bits. Figure
3.1 illustrates the main and basic operations for the generations and applications of
one or more big memorizable secret(s).
3.6 Potential Applications of Available Big Memorizable Secret
With the realization of big memorizable secret, not only the big secret keys of
symmetric key crypto systems of higher security strength like AES-192 and AES-256
can be realized firstly, but memorizable public-key crypto system (MePKC) secondly,
and other cryptographic, information-hiding, and non-cryptographic applications
thirdly, in the field of information engineering that need big and yet memorizable
secret.
27
100
V
User selects one or a mixture of the methods as follows to create
one or more big memorizable secrets in a computing device:
(1) Self-created signature-like Han character of CLPW & CLPP
(2) Two-dimensional key (2D key)
(3) Multilingual key
(4) Multi-tier geo-image key
(5) Multi-factor key using software token
101
102
The created secret is used as password, passcode (aka pin), symmetric key,
asymmetric private key, stego-key, symmetric watermarking key, asymmetric
watermarking private key, PRNG seed, etc., for one or a mixed combination of the
systems as follows in the field oi information engineering:
(1) Cryptographic applications like 256-bit AES, DSA， ECC， MePKC
(2) Information-hiding applications like steganography, watermarking
(3) Non-cryptographic applications like PRNG, CSPRBG
V
Perform one of the many functions as follows:
(1) Creating an asymmetric public key using an asymmetric private key
(2) Encrypting using a symmetric key, stego-key
(3) Decrypting using a symmetric key, stego-key, asymmetric private key
(4) Signing using an asymmetric private key
(5) Embedding using a symmetric watermarjang key, asymmetric WM private key
(6) Verifying using a symmetric watermarking key
(7) Creating an HMAC (Keyed-Hash Message Authentication Code) using a secret key
(8) Seeding PRNG, CSPRBG
(9) Other functions using secret(s)
After finisning the process using the secret, do either one of the
processes as follows before the application is closed:
(1) Delete the secret immediately during or after the application
(2) Store the secret for limited time
(3) Store the secret for limited amount of usages
(4) Store the secret for limited amount of usages per unit of time
/X7l
104
Figure 3.1 Generations and applications of one/more big memorizable secrets
28
These cryptographic applications include cryptographic schemes like
encryption, signature, key exchange, authentication, blind signature, multisignature,
group-oriented signature, undeniable signature, threshold signature, rail- stop
signature, group signature, proxy signature, signcryption, forward- secure signature,
designated- verifier signature, public-key certificate (aka digital certificate), digital
timestamping, copy protection, software licensing, digital cheque (aka electronic
cheque), electronic cash, electronic voting, BAP (Byzantine Agreement Protocol),
electronic commerce, MAC (Message Authentication Code), key escrow, online
verification of credit card, multihash signature, etc.
Those information-hiding applications include steganographic and
watermarking schemes like stego-key in steganography, secret key in symmetric
watermarking, private key in asymmetric watermarking, etc. Meanwhile, the non-
cryptographic applications are PRNG (Pseudo-Random Number Generator) and
CSPRBG (Cryptographically Secure Pseudo-Random Bit Generator). Hence, there
exist lots of needs to have big memorizable secret for lots of cryptographic,
information- hiding, and non-cryptographic applications in the field of information
engineering, generally, and security engineering, particularly.
3.7 Future Development of Keys the Secret
These keys the secret need good generation methods (Scalet, 2005) and key
management (Fumy & Landrock, 1993; Beach, 2001; Witty, 2001). Wailgum (2008)
questioned on whether there were too many passwords or humans were lacking of
memory power. In term of memory, there are two forms: Recognition-based and
recall-based. Weinshall and Kirkpatrick (2004) presented those recall-based
passwords. Bill Gates with Microsoft has once claimed the ending of the passwords
(Allan, 2004; Kotadia, 2004; Fried & Evers, 2006).
Subsequently, there are introductions of some password alternatives like
Information Card (Wikipedia Contributors, 2008ap), Windows CardSpace (Wilson,
2008)， Higgins Project, OpenID (Wikipedia Contributors, 2008am), Identity
Metasystem (Jones, 2005; Cameron & Jones, 2006)， Identity Selector, digital identity
29
(Cameron, 2005; Cavoukian, 2006; Wikipedia, Contributors, 2008al)， Authorization
Certificate, Extended Validation Certificate, etc.
Furnell (2005) analyzed whether human could get rid of passwords and
concluded that passwords could not be replaced. Here, if the inventions and
innovations on big secret(s) creation methods and their applications are adopted,
especially MePKC (Memorizable Public-Key Cryptography), the complicated
mentioned password alternatives may be made simpler or at best be avoided. More
literatures on password are available at PasswordResearch.com website [URL:
www.passwordresearch.com] .
For security of asymmetric key crypto systems, the mathematical hard
problems depend on the researchers' creativity and innovation as well as the
computing technologies to crack them. For example, the cryptanalytic attacks like
Wiener (1990) and so on, that can be discovered in the future, may request for longer
asymmetric key sizes and/or other mathematical hard problems. Challenges with
awards offered by the PKC services providers to crack certain PKC with certain key
sizes are always there for the public to attempt.
Anyway, the practically secure key sizes for symmetric and asymmetric key
crypto systems at different protection periods are always under the regular
evaluations by a lot of researchers (Williams, 2002). Website of KeyLength.com
[URL: www.keylength.com] ("Cryptographic Key Length Recommendation," No
date) is a collection database for lots of documentations on these practically secure
key sizes for various applications in security engineering, particularly, and
information engineering, generally.
30
CHAPTER 4 CREATING BIG MEMORIZABLE SECRET
(PART 2)
4.1 Passphrase Generation Methods
Civilian crypto system applies Kerckhoff s law to have security dependency
100% on the password secrecy. This reflects the fact that key length and key space
are very important to ensure enough entropy or randomness to secure a crypto system.
For stronger password, passphrase is suggested. Currently, there are three methods to
generate passphrase: Acronym, full sentence and diceware. Here, an alternate method
to diceware is proposed: Coinware (Lee & Ewe, 2006)， by using the coin. Coinware
uses four coins to generate one hexadecimal digit. The created word lists are in
hexadecimal order and can be applied for multilingual passphrase generation. Its
exemplary application for Chinese language password is then shown. Readily-made
Chinese character word list in the Unicode CJK unified ideographs enables fast
hexadecimal reading for random passphrase generation. Hanyu Pinyin and Sijiao
Haoma are used to Romanize and uniquely represent each Han character. Jyutping
and Romaji are then used for Cantonese and Japanese languages, respectively.
Table 4. 1 Passphrase generation from acronym
Sentence
Passphrase
Passwords should be impossible to remember and never written down
psbitranwd
Passwords should be impossible to remember and never written down
PsBiTrAnWd
Good or bad, you have to do it.
Goby,htdi.
Good or bad, you have to do it.
Drd,ueoot.
It may be a few sentences. One, two or more.
Imbafs.O，tom.
4.1.1 Acronym
For the passphrase created using the acronym (Schneier, 1996; PGP
Corporation, 2006; Yan, Blackwell, Anderson & Grant, 2004)， a user has to
remember one or a few sentences. Then, the first, second, or last, etc. characters of
each word in the sentence(s) are joined to form an acronym. Both alphanumeric and
31
non-alphanumeric ASCII characters may become the character of the acronym. The
techniques of capitalization and permutation may be used to increase the randomness.
This acronym will then act as the key. It has the features of high randomness and
short key length. The examples of this method are in Table 4.1.
4.1.2 Full Sentence
The passphrase generation using the acronym is sufficient if the key length
requirement is short. When the minimum key size demand is long, normally one full
sentence or a few short sentences are entered directly as the key (Schneier, 1996;
PGP Corporation, 2006; Yan, Blackwell, Anderson & Grant, 2004). So far, it is an
open problem to type the entire phrase into a computer with the echo turned off
(Schneier, 1996). If the masked password is shown during the password entering
process, then it will subject to shoulder surfing attack.
Besides, since the passphrase of full sentence has each word to be selected
associatively, its randomness is magnitude-wise high but relatively low if password
ciphertext is available. For example, superuser of any computing system can easily
obtain ciphertext of the password. By gaining access to the encrypted password, the
threats of ciphertext-only attack and frequency analysis of short cryptogram (Hart,
1994; Lee, Teh & Tan, 2006) are then possible. For instance, the unicity distance of
English language is about 30 characters. Once the encrypted password is equal to or
more than the unicity distance, unique decipherability of the encrypted password will
be feasible.
4.1.3 Diceware
Using full sentence for passphrase generation, the word frequency
distribution can be under computational analysis (Kucera & Francis, 1967). To get
rid of the association of words, diceware (PGP Corporation, 1996) introduced by A.
G. Reinhold is an improved passphrase generation method.
There are many software pseudo-random number generators (PRNGs).
Unfortunately, they have lots of pitfalls (Eastlake, Crocker & Schiller, 1994) to ease
32
any possible attack. Hence, some hardware random number generators (RNGs) such
as coin and dice are very much better than the software PRNGs.
Diceware uses dice to select a word from an ordered word list. The word list
can be in any language and based on senary or base-6 numeral system. For the most
popular diceware, it is an English word list with 7776 (= 6 5 ) words. Five dice values
are needed to locate one word randomly. Every selected word carries entropy of
12.92 bits. Table 4.2 shows the minimum diceware words for different security levels.
Table 4.2 Minimum diceware words (7776 word list) for different security levels
Key Size (bit)
32
64
72
80
112
128
256
Diceware
word
3
6
7
9
10
20
bit
38
64
77
90
116
129
258
4.1.4 Coinware
In addition to diceware using dice, coinware using coin is proposed here.
Coin tossing is conducted to generate random passphrase. Each face of the coin is
labeled as binary bits "0" and "1"， respectively. Four coin values are used to derive a
hexadecimal digit. Therefore, the word list is in hexadecimal order. Table 4.3 shows
the conversions between the binary (BIN) and hexadecimal (HEX) numeral systems.
Table 4.3 Conversions between binary and hexadecimal numeral systems
BIN
HEX
BIN
HEX
BIN
HEX
BIN
HEX
0000
0100
4
1000
8
1100
C
0001
1
0101
1001
9
1101
D
0010
2
0110
6
1010
A
1110
E
0011
3
0111
1011
B
1111
F
4.1.5 Monolingual, Bilingual, and Multilingual Word Lists
Having word list and random number generator, computational analysis on
word frequency distribution is avoided and random passphrase generation is ensured.
For the word list, one may use readily made word list or prepare a new word list.
33
For readily made word list, it is normally monolingual unless one combines
two or more monolingual word lists with different languages. To prepare a new word
list, one may go for monolingual, bilingual, or multilingual to suit one's linguistic
ability. The purpose to have word list consisting of more than one language is to
increase the key space and consequently the key entropy per word.
For the word list, each word has to be unique, short, and memorizable. Start
with the shortest word. Then slowly increase the character length of the word until
the key space setting of the word list is met. To be easy, one may set the key space of
monolingual word list to 4096 (= 2 12 ) or 8192 (= 2 13 ) words. Two or more
monolingual word lists can be joined to form bilingual or multilingual word lists.
Nowadays, when language proficiency is excluded, bilingual or multilingual
people ("Mother Tongue," No date; Paradowski, No date; Wikipedia Contributors,
2008p， 2008aa， 2008ab， 2008ae， 2008av) exceed the monolingual people (Wikipedia
Contributors, 2008a), where English language is a second language due to its soft
power of computer language in the recent decades and hard power of British Empire
and American influence since the 19 th century (Wikipedia Contributors, 2008bb，
2008bc). The language studies (Matthews, 1997; Crystal, 1999; Finegan, 2004;
Gordon, 2005; Wikipedia Contributors, 2008ai) are based on text corpus (Wikipedia
Contributors, 2008z)， branching into a few fields like theoretical linguistics, applied
linguistics, and corpus linguistics (Wikipedia Contributors, 2008h， 2008u， 2008y).
The language policy (Wikipedia Contributors, 200 8j) of a country has
determined the language(s) in education and the speakers of a language. There are
various languages spoken and written by various people from different countries
(Wikipedia Contributors, 2008c, 2008m, 2008p, 2008ac， 2008ad， 2008aq, 2008ar，
2008au). For example, there is a statistical analysis of Chinese language (Zhang, Xu
& Huang, 2000). Till here, we can see that the bilingual and multilingual word lists
have their demands and can increase the key space of coinware for higher complexity.
4.1.6 Key Length Requirements of Coinware
It is important to know the key size equivalence for symmetric and
asymmetric (RSA, discrete logarithm and elliptic curve) crypto systems (Schneier,
34
1996; Williams, 2002; Gehrmann & Naslund, 2005， 2006， 2007) for different
security levels. Tms step enables a user to prepare suitable and sufficiently strong
password or passphrase before opening an account and conducting an encryption.
Table 4.4 shows this important information (Williams, 2002).
Table 4.4 Key size equivalence for symmetric and asymmetric crypto systems (bit)
Symmetric Key Size (bit)
48
64
80
112
128
192
256
RSA
480
816
1248
2432
3248
7936
15424
Discrete
Logarithm
Field Size
480
816
1248
2432
3248
7936
15424
Subfield
96
128
160
224
256
384
512
Elliptic Curve
96
128
160
224
256
384
512
The minimum coinware word relies on the key space of the word list. The
monolingual, bilingual, or multilingual word lists of 4096， 8192， 12288， 16384， and
24576 words have entropies per word of 12, 13， 13.58, 14, and 14.58 bits,
respectively. Table 4.5 shows the minimum coinware words for various word list
sizes (WLS).
Table 4.5 Minimum coinware words for various word list sizes (WLS)
Symmetric Key Size (bit)
48
64
80
112
128
192
256
WLS 4096
word
4
6
7
10
11
16
22
bit
48
72
84
120
132
192
264
WLS 8192
word
4
7
9
10
15
20
bit
52
65
91
117
130
195
260
Coinware
WLS 12288
word
4
6
9
10
15
19
bit
54
67
81
122
135
203
258
WLS 16384
word
4
6
8
10
14
19
bit
56
70
84
112
140
196
266
WLS 24576
word
4
6
8
9
14
18
bit
58
72
87
116
131
204
262
The current common demands of security levels are 80- and 128-bit for the
symmetric crypto system. These security levels ensure protection of 5 and 30 years,
respectively. From Table 4.5, word list size of 8192 is suitable for monolingual and
35
bilingual users. Monolingual users can use a monolingual word list of 8192 words.
Meanwhile, bilingual users can use two unique monolingual word lists of 4096
words each. For multilingual users, word list size of 24576 is suggested where three
unique monolingual word lists of 8192 words each can be used.
4.2 Chinese-Character-Encoded Passphrase
For coinware applications, there are readily-made word lists in Unicode (The
Unicode Consortium, 2006) for various languages. This is because both coinware
and Unicode are in the hexadecimal order for their word lists. This is especially true
for the CJK languages of Chinese language, Japanese language, and Korean language
that use the Han characters. Word list is also a character list for CJK languages. Here,
we discuss on the Chinese language password generation with optional coinware.
In computing system, password is dominated by Roman alphabet or Latin
alphabet due to character encoding of ASCII. Here, a pronounceable and
memorizable password policy in Chinese language is proposed. Phonetic encoding of
Hanyu Pinyin (Popular Book, 2003) and symbolic encoding of Sijiao Haoma (or four
corner method) (Wikipedia Contributors, 2007a) are used to create the uniqueness of
each Chinese character or Han character for alphanumeric representation.
Based on about 70229 Han characters in the Unihan database of CJK
ideographs for the version of Unicode 4.1, each Han character has entropy of 16.1
bits. Five Han characters will satisfy the 80-bit minimum randomness requirement of
symmetric key crypto system for strong password. Self-created signature-like Han
character and passphrase represented by printable ASCII generates shorter, stronger
and more memorizable passwords with 27.3 and 85.4 bits per Han character,
respectively. Other CJK languages using the Han characters like Cantonese language
and Japanese language are also applicable via the pronunciation Romanization
systems of jyutping and romaji, respectively.
Han character is also called Chinese character (Wikipedia Contributors,
2007h). For the Chinese character input methods, they are either based on
pronunciation, character structure, or a combination of pronunciation and character
structure (Wikipedia Contributors, 2007c). These methods are closely linked to
36
Chinese character encodings (Wikipedia Contributors, 2007d) to allow a user to enter
a Chinese character. A user normally remembers the pronunciation and/or character
structure of a Chinese character to facilitate its input.
Huang (1985) proposed a type of Chinese character encoding to ease Chinese
input by using the combination of pronunciation and character structure. This kind of
Chinese character encoding can create a key per Chinese character. The maximum
size of this encoding is six characters, where there are three characters for phonetic
sound, one character for tone, and two characters for character structure. The
memorizability of this Chinese-character-encoded key is better than the Environ
password, but its security is subject to dictionary attack.
4.2.1 Environ Password
An analogue to the Romanization of Chinese language to have alphabets and
digits is the Environ password (Anderson, 2001， p. 49). Good memorizability exists
when it is linked to a learnt language. For English language, U.K. government
introduced the case insensitive Environ password in October 2005 for short-term
protection (Wikipedia Contributors, 2007b). It has an 8-character key pattern as in
Table 4.6. This pronounceable password has 34.9 bits per unit.
Table 4.6 Environ password
Form
[consonant - vowel - consonant - consonant - vowel - consonant - digit - digit]
[consonant - vowel - consonant - digit - consonant - vowel - consonant - digit]
Example
pinray34, yankan77， supjey56， kinkin99; pin3ray4, yan7kan7, sup5jey6， kin9kin9
4.2.2 Unicode
Unicode unifies the Han characters of CJK languages into CJK unified
ideographs or Unihan under ISO 10646. There are three major blocks of Han
characters or Chinese characters in the Unicode character encoding: CJK unified
ideographs, CJK unified ideographs extension A, and CJK unified ideographs
extensions B. For the mean time, Unicode Consortium is preparing the CJK unified
ideographs extension C and CJK unified ideographs extension D (Wikipedia
37
Contributors, 2008at). The CJK unified ideographs extension C with 4251 Han
characters will be included into the next version after Unicode 5.1.
For Unicode 4.1， the first block lists the Han characters from [4E00] to
[9FBB] in hexadecimal value. The second block lists from [3400] to [4DB5]. The
third block lists from [20000] to [2A6D6]. Hence, there are three readily made word
lists or character lists for Chinese language. These word lists have 20924， 6582 and
42711 words or characters, respectively. In addition, there are CJK compatibility
ideographs having 12 characters. For a combined word list, it is a key space of 70229
characters. After radical exclusion, the key space has about 70000 characters. This
forms a Chinese language word list with high entropy of 16.10 bits per Han character.
To start coinware, first flip or toss a coin to randomly select a binary bit "0"
or "1". If bit "0"， the first and second blocks of CJK unified ideographs and CJK
unified ideographs extension A are chosen. If bit "1"， the third CJK block of CJK
unified ideographs extension B is chosen. Then continue with coin tossing to obtain
four coin values representing four binary bits. These four binary bits are converted
into one hexadecimal digit. Repeat coin tossing to get four coin values for another
three rounds. Four randomly obtained hexadecimal digits will locate the unique Han
characters in the previously selected CJK block(s). These three blocks are available
at [URL: http://www.unicode.org/charts/]. If the hexadecimal digits do not hit any
Han character, get another set of hexadecimal digits. Coming to here, the selected
Han character will need Chinese character Romanization to enable computer input.
4.2.3 Chinese Language Password (CLPW)
Zhonghua Zihai in 1994 has 85,568 Chinese characters (Wikipedia
Contributors, 2007d). It means a Chinese character may have entropy of 16.38 bits
when the Unicode Unihan database is further enlarged. For key security, this is an
advantage over the 6.57-bit ASCII characters, which are used for the Latin languages.
For computers with support of Chinese character encoding, Chinese language
password (CLPW) is shorter for the similar key size of ASCII-based password. This
indicates better memorizability. For computers without support of Chinese character
38
encoding, which are general for majority of the computers, Romanization of Chinese
language is needed to create the same advantage in term of memorizability.
Chinese input methods and Chinese character encodings can be used to
Romanized CLPW. The Romanization of Chinese language is either based on
pronunciation, character structure, or a combination of the both. To uniquely
represent a Chinese character, Huang (1985) has a good reference for CLPW
Romanization, where both Huang pronunciation and character structure are used to
create a Chinese-character-encoded word with a maximum of 6 characters.
However, this approach requires modernization. The pronunciation system of
Hanyu Pinyin (汉 语拼" ir) (Popular Book, 2003) and character structure system of
Sijiao Haoma or four-comer method ( 四角 号 石马) (United Publishing House, 2001,
2002; Wikipedia Contributors, 2007a) are proposed to create a Chinese language
password. In Hanyu Pinyin, there are 415 unique syllables with 22 initials (or onsets)
and 39 finals. This pronunciation system is illustrated in Table 4.7.
Table 4.7 Phonetic encoding of Hanyu Pinyin (Mandarin-based)
Initial
(22)
nil
b
P
m
f
d
t
n
1
g
k
h
j
q
X
z
c
s
zh
ch
sh
Final
(39)
a
e
ai
ei
ao
ou
an
en
ang
eng
ong
ia
io
ie
iao
iu
ian
in
iang
ing
iong
u
ua
uo
uai
ui
uan
un
uang
ueng
ii
iie
iian
iin
er
N.B.: For Romanization, e and u can be represented by [oe] and [v], respectively.
In addition to initials and finals, there are 5 tone marks. These tone marks are
numbered as 1, 2, 3, 4 and d in corresponding with Yinping (阴 平) ， Yangping
( 阳^ F ) ， Shangsheng (上 尸ノ ， Qusheng ( 去声 ) and Qingsheng ( 轻声 ) .
Then, the 4+1 -digit Sijiao Haoma is added to describe the character structure
of a Chinese character. The upper left number is the first digit. The upper right
39
number is the second digit. The lower left number is the third digit. The lower right
number is the fourth digit. The fifth digit is Fuhao or attached number (附 号) ，
which represents the middle character structure on the right hand side. Figure 4.1
shows a Cmnese poem to easily memorize the Sijiao Haoma. Table 4.8 shows the
strokes represented by Sijiao Haoma.
横 ᅳ 垂 二三 点 掠
叉 四插五 方框六
七 角 八 八 九是小
点 下 有 横 变零头
Figure 4.1 Chinese poem ror easy memorization or Smao Haoma
Table 4.8 Character structure encoding of Sijiao Haoma
Stroke name (笔 名)
Digit (号 码)
Stroke (笔 形)
Tou (头)
、 .
Heng (横)
1
Chui (垂)
2
I i J
Dian ( 点 )
3
Cha (叉)
4
十 J、
Chuan ( 串 )
^ 丰
Fang ( 方 )
6
P □
Jiao (角)
ᄀ 厂
Ba (八)
8
A 入 入
Xiao (/]、)
9
小 十
lable 4.9 Forms of Romanized Chinese-character-encoded words
Form
[Hanyu Pinyin] (Tone Mark) [Sijiao Haoma] (Fuhao)
Example
han3714
han43714
han37140
3714han
3H7A1N4
Finally, the Hanyu Pinyin, tone mark, Sijiao Haoma, and Fuhao are joined to
form a Romanized Chinese-character-encoded word as in Table 4.9 for the Chinese
character of Han (汉) . Tone mark and Fuhao are optionally included. For the
computer input or cLPW for Han character, the Hanyu Pinyin and Sijiao Haoma can
40
be typed side by side. If the coin tossing gives a 5-digit hexadecimal string of
[06C49], then the Han character of (汉) is selected from the Unicode. Table 4.9
shows the possible forms for the Chinese character Romanization of (汉) [Hanyu
Pinyin = han4] [Sijiao Haoma = 37140].
This creates a Chinese-character-encoded word ranging from 5 to 12 ASCII
characters. Several Chinese-character-encoded words can be used as a Chinese
language password. The capitalization and permutation can slightly increase the
entropy. However, it is suoject to dictionary attack.
4.2.4 Key Length Requirements for Coinware in Chinese Language
Referring to the created Han character combined list in Section 4.1, it is 16.10
bits per Han character. Using this key entropy, minimum coinware words at different
security levels for Chinese language password can be derived as in Table 4.10.
Due to high key entropy of Chinese language word list, it is obvious to
observe the significant drop of minimum coinware words as from Tables 4.5 and
4.10. As Japanese language and Korean language are using the Han characters as
well, similar word lists with large key space can be created. For Chinese language
family, it is applicable to Mandarin, Wu, Cantonese, Min, Jin, Xiang, Hakka, Gan,
Hui, and Ping languages/dialects with speaking population of 800， 90， 80， 50, 45， 35,
35， 20， 3， and 0.2 millions, respectively.
Table 4.10 Minimum coinware words for Han character combined list (70229 words)
Symmetric Key Size (bit)
48
64
80
112
128
192
256
Coinware
(Chinese Language)
word
3
4
5
8
12
16
bit
48
64
80
112
128
193
257
4.2.5 Self-Created Signature-like Han Character
The combination of 415 Hanyu Pinyin syllables, 5 tone marks, and 10,000
Sijiao Haoma numbers are more than enough to encode all the Han characters
41
available at present. In order to increase the randomness or entropy of Han character,
the creation of new Han character is a must.
This situation happens in real life for the individual name in gaining
uniqueness. The created Han character is also signature-like. For Han character
creation, it may follow the six methods of Liushu (六 书) (Huang, 2002). The
Liushu includes Xiangxing (象形) (picto grams), Zhishi (指 事) (ideograph),
Huiyi (会意) (logical aggregates), Xingsheng (形 尸ノ (pictophonetic
compounds), Jiajie (假 借) (borrowing), and Zhuanzhu (转汪 ) (associate
transformation) (Wikipedia Contributors, 2007h; Xu， 2001; Luo, 1990， 2003).
Figure 4.2 Example of self-createa signature-like Han character by modifying the
Han character of (汉) from [Hanyu Pinyin = han4] and [Sijiao Haoma = 37140] to
[Hanyu Pinyin = han4] and [Sijiao Haoma = 37141]
An example of created Han character is shown in Figure 4.2. The Han
character of (汉) is modified from [han437140] to [han437141] by adding a
horizontal stroke between the upper right corner and lower right corner. Self-created
signature-like Han characters enlarge the key space of Chinese language password to
4,150,000. When tone mark and Fuhao are included, it becomes 207,500,000 or
entropy of 27.63 bits per Han character. The efficiency of Chinese language
password is greatly increased.
4.2.6 Self-Created Chinese Language Passphrase (CLPP)
To further increase the entropy of Chinese language password, we can have
self-created Chinese language passphrase (CLPP). At least one non-alphanumeric
character has to be included together with capitalization, permutation, character
stuffing, and text-based semantic noises. Character stuffing is like bit stuffing in data
42
communication to enable the syllable length at a fixed value of 6. It is 6 because the
maximum syllable length is 6， excluding the tone mark.
Adding fixed syllable length, tone mark, Sijiao Haoma with Fuhao, and one
non-alphanumeric character together, a string of 13 ASCII characters is obtained as a
basic unit of a self-created CLPP. The non-alphanumeric character is used as a
separator and text-based semantic noise.
Table 4.10 shows examples of self- created CLPP. This Chinese-character-
encoded passphrase has 85.41 bits per Han character. It has good memorizability,
resistance to dictionary attack, and suitability for general password usages.
Table 4.11 Forms of self-created Chinese language passphrase for ( 汉 )
Form
No character stuffing
With character
stuffing & noise
Capitalization &
permutation
Example
han4&37140
h@n4«* & 37140
37140&HaN4*"
4.2.7 Cantonese Language Password Using Jyutping
Han unification of Unicode builds Han characters database for CJK
languages (Chinese, Japanese, and Korean). The proposed password and passphrase
generation method can be applied to any CJK languages using the Han characters by
changing the pronunciation Romanization system. The character structure encoding
of Sijiao Haoma remains the same for all the Han characters in any CJK languages.
Cantonese language is used by a global population of about 80 millions.
Being the official language in Hong Kong SAR (Special Administrative Region) and
Macau SAR of PRC (People's Republic of China), the regulation works of
Cantonese language are done here. It shares majority of the Han characters with
Chinese language in Mandarin except those Han characters in the HKSCS (Hong
Kong Supplementary Character Set). For HKSCS-2004, it has 4941 Han characters
as in year 2004 under ISO 10646 standard. Hence, it is compatible with Unicode,
which implements the ISO 10646 standard.
43
There are many Cantonese pronunciation systems. Among them, two systems
are Romanized and computer friendly. One of them is the standard Cantonese pinyin
or HKED ( 《常用 字廣州 話讀音 表》 拼音方 案) ( 「教院 式」 拼音方 案) .
This is the only pronunciation Romamzation svstem accented bv Education and
Manpower Bureau of Hong Kong and Hong Kong Examinations and Assessment
Authority. Another is jyutping proposed by LSHK (The Linguistic Society of Hong
Kong) in year 1993.
Nowadays, regulation works of Cantonese pronunciation for Unicode adopt
jyutping system. Han characters in Unicode are matched with jyutping, where the
lists are downloadable from the URLs of [http://www.isol0646hk.net/jp/index.jsp]
and [http://www.info.gov.hk/digital21/eng/structure/jyutping.html].
Jyutping can also be applied into the coin tossing of coinware. As for the
hexadecimal strings of [03400]， [04E00], and [0E000], the jyutping for these
Cantonese characters are [jaul], [jatl], and [mou5], respectively.
Table 4.12 Phonetic encoding of jyutping in Cantonese language
nil
b
P
m
f
d
t
n
1
Initial (20)
g
k
ng
h
gw
kw
w
z
c
s
j
ip
it
ik
im
in
ing
iu
yu
yut
yun
u
up
ut
uk
um
un
ung
ui
e
ep
et
ek
em
en
eng
ei
eu
Final (59)
eot
eon
eoi
oe
oet
oek
oeng
ot
ok
on
ong
oi
ou
ap
at
ak
am
an
ang
ai
au
aa
aap
aat
aak
aam
aan
aang
aai
aau
In jyutping system, there are 20 initials and 59 finals as in Table 4.12. These
initials and finals construct about 629 syllables for Cantonese language as compared
44
to 415 syllables for Chinese language in Mandarin. For tone mark, 6 distinct tone
contours are used for 9 tones. For completeness, the jyutping has syllables that have
no matching Han character. Nevertheless, in the application for Cantonese language
password, all jyutping syllables may be useful for self-created password.
Table 4.13 shows the examples or Cantonese language password. It is similar
to Chinese language password in Mandarin. Sijiao Haoma is exactly encoded. For
jyutping, the maximum syllable length is 6. Capitalization, permutation, and
character stuffing can be used to generate self-created signature-like Cantonese
language password and passphrase. The key space of self-created Han characters in
Cantonese can reach 377,400,000 keys or 28.49 bits per Han characters.
Table 4.13 Forms of self-created Cantonese language passphrase for (汉)
Form
Traditional Chinese
(漢)
Simplified Chinese
(汉)
With character stuffing
Example
hon3&34185
hon3&37140
hon3*** & 34185
4.2.8 Japanese Language Password Using Romaji
In Japanese language, there are four writing systems: Two syllabaries of
Hiragana (平仮名) and katakana (片仮名)， one logogram of kanji (漢字)， and one
Romanization of romaji (ローマ字). The most widely used Hepburn Romanization
is adopted for romaji. The password generation method for Chinese language
password can be used for Japanese kanji via the combination of romaji and Sijiao
Haoma.
Firstly, obtain the Sijiao Haoma with Fuhao for the Japanese word in kanji.
Then, the kanji is converted to romaji for pronunciation Romanization. Character
stuffing is longer for Japanese language password as the kanji is having variable
number of syllables from a minimum of one syllable. For Hepburn Romanization,
there are about 132 syllables.
Table 4.14 shows examples of kanji passwords. To avoid dictionary attack,
self-created kanji with character stuffing, capitalization, and permutation, as in
45
Section 4.2.5 can be used. Coinware allows random selection of Han characters (Lee
& Ewe, 2006).
Table 4.14 Forms of Japanese language password for (大 ) ， (漢 ) ， and (山)
Form
dai (大) (だい)
kan (漢) (かん)
yama (山) (や ま )
Example
dai&40800
kan&34185
yama&22770
4.2.9 Key Length Requirements for Various Chinese Key Spaces
For unbreakable encryption, the key size has to be at least the same with
message size as in one-time password (Shannon, 1949). So far, the full Unihan
database of 70229 Han ideographs in Unicode 4.1 is used to build the Chinese
language password and passphrase. In the Han unification of Unicode, Han
ideographs are called as Hanzi in Chinese language, Kanji in Japanese language, and
Hanja in Korean language.
If only the basic Unihan database of 20948 Han ideographs in the CJK
unified ideographs are used, by excluding the CJK unified ideographs extension A
and CJK unified ideographs extension B， the entropy will drop from 16.10 to 14.35
bits per Han character. Then more Han characters are required to fulfil the key length
requirements. Hence for short, strong, and memorizable password, self-created
signature-like Chinese language password and passphrase are in favourite. The
situation of various Chinese key spaces is shown in Table 4.15.
Table 4.15 Minimum key lengths for various Chinese key spaces (in Han character)
Database
Key Space
Entropy
(bit 1 Han char.)
Minimum Key Length
(in Han character)
80-bit
128-
bit
192-
bit
256-
bit
Basic Unihan
20948
14.35
6
9
14
18
Full Unihan
70229
16.10
5
8
12
16
Self-created Han Character
166,000,000
27.31
3
5
8
10
Self-created Cantonese
377,400,000
28.49
3
5
7
9
Self-created Passphrase
95 13
85.41
1
2
3
3
46
4.2.10 Example to Create CLPW and CLPP
This section shows an example on how self- created signature-like Han
character is encoded to create big and yet memorizable secret. For the word
etymology of "Chin" and "Han", they are originated from the names of two early
dynasties called Qin (秦) and Han (漢) in China. Even though there are many
rounds of renaming in Chinese language for the country of China, its English name
remains unchanged till today in carrying the phoneme of "Qin" for "Chin". Therefore,
Chinese character is also called Han character (漢子 or 汉子) . The repertoire size
of Han characters is 85,568 in the dictionary of Zhonghua Zihai (Word Dictionary of
Chinese Language) published in 1994. Han characters are used in CJKV languages,
in which it is called Hanzi in the Chinese language, Kanji in the Japanese language,
Hanja in the Korean language, and Han Tu in the Vietnamese language.
It is to note that the entropy of Han characters is higher than the ASCII
characters. Due to the logographic type of language, Han characters carry visual
meaning and hence are easily memorizable. In other words, Han characters have the
intrinsic features of high entropy and good memorizability, which mean their
suitability for the creation of big and yet memorizable secret. Nevertheless, Han
characters have input problem. The number of Han characters is too many to be
represented by a single keyboard. Another problem is that direct application of Han
characters as password the secret is vulnerable to guessing attack, dictionary attack,
and pre-computation attack.
To solve the first problem, a Han character can be encoded using its character
structure (or symbolic shape) and/or phonetic pronunciation based on ASCII
characters. This process is called Romanization. For example, when pronunciation
system of Hanyu Pinyin (汉 语拼 首ノ and character structure system of Sijiao
Haoma (or four-corner method) ( 四角 号石马 ) are used to encode and Romanize the
Han character of {han} (、汉 ) in simplified form, the code is {han4} from Hanyu
Pinyin and {37140} from Sijiao Haoma, forming one of many possible codes like
{han437140} called CLPW (Chinese Language Password). However, the second
47
problem of vulnerability to guessing attack, dictionary attack, and pre-computation
attack, has not yet been solved.
To solve the second problem, the randomness of the CLPW using Han
character has to be increased. A Han character from any encoding like Unicode
encoding can be modified to become a self-created signature-like Han character new
to the current available repertoire of Han characters. Phonetic pronunciation system
and character structure system using ASCII characters can be used to encode and
Romanize the self- created signature-like Han character into a CLPW that can resist
the guessing attack and dictionary attack.
200
201
Figure 4.3 Example of self-created signature-like Han character from {han} (汉)
Figure 4.3 illustrates an example of self- created signature-like Han character
by modifying the Han character of {han} (汉) in Box 200 of Figure 4.3 (left) from
{hanyu pinyin = han4} and {sijiao haoma = 37140} to Box 201 of Figure 4.3 (right)
{hanyu pinyin = han4} and {sijiao haoma = 37141}. In other words, the CLPW has
been modified from {han437140} to {han437141}. The adoption of self-created
signature-like Han character shares the similar habit with Chinese people to use a
general name aliasing with another rare name. A name using frequently used Chinese
characters allows easier memorizability and pronunciation, but harder differentiation
due to name clashing. A second alias name using rarely used Chinese characters
helps to make a person's name unique and differentiable from the others, but carries
a problem of harder pronunciation. Hence, pronounceable name is for easy calling
and unique name is for easy differentiation.
48
Self-created signature-like Han characters enlarge the key space of CLPW to
4,150,000. When tone mark and Fuhao (附 号) are included, it becomes
207,500,000 or entropy of 27.63 bits per Han character. The efficiency of CLPW is
hence greatly increased. To further increase the randomness, a Chinese language
password (CLPW) can be upgraded to a Chinese language passphrase (CLPP) by
adding textual semantic noises like character stuffing, capitalization, permutation,
punctuation marks, misspelling, mnemonic substitution, and/or alternative symbols
from ASCII mutual substitution table. One unit of CLPW can be set to a fixed length
like 13 ASCII characters or other size, and a few units of CLPW form a unit of CLPP.
For a unit of CLPW, its 13 ASCII characters are formed from phonetic syllable of
length 6， tone mark of length 1， Syiao Haoma with Fuhao of length 5， and non-
alphanumeric character as a separator of length 1.
Character stuffing is like bit stuffing in data communication to enable the
syllable length at a fixed value of 6. It is 6 because the maximum syllable length is 6
in Hanyu Pinyin, by excluding the tone mark. Of course, other phonetic
pronunciation systems, especially Chinese dialects and CJKV languages, like
jyutping for Cantonese language and romaji for Japanese language, can be used as
well. Similarly, other encodings of Han characters could be used. For the example of
13-character CLPW with textual semantic noises using the Han character of {han}
(汉) ， it can be in the forms of {h@n4*** & 37140}， {37140&HaN4***}， and so
on. When the textual semantic noises are good enough from prediction, the ideal
entropy of fully random absolute rate at entropy of 85.41 bits per unit of CLPW (or
unit of Han character with modification and added noises) can be approached.
A few serial units of CLPW form a CLPP that has good memorizability,
resistance to guessing attack and dictionary attack, as well as suitability for general
usages. CLPP of size beyond 128 bits can realize the AES-128, AES-192, AES-256,
DSA-256, ECC-256, and so on. When CLPP is used for MePKC operating on the
platforms of FFC and ECC, even the pre-computation attack can be avoided. Table
3.3 shows the numbers of CLPP units for various key sizes. People knowing Han
characters can memorize a CLPP with 2 to 4 units of CLPW as easy as remembering
a person's name using rarely used Han characters.
49
Nevertheless, the current prior art of single-line key/password input neld is
not that friendly when there are two or more CLPW. There exists a user interface
problem to input password with long key size in a single line. This problem happens
also to other passphrases having a lot of characters. Whenever there is a pause or
interrupt during the input process of a passphrase, it is hard to determine the starting
points of every word or unit of a passphrase. In other words, a long passphrase like
three to four units of CLPP has to be entered instantly without an interrupt or error.
Any uncertainty in keying in a passphrase to a single-line key field requires the
whole re-keying process of that passphrase.
4.3 Two-Dimensional (2D) Key
Conventionally, single-line key field is used to input a key. The selection of a
key depends on the factors of memorizability and security. The minimum key sizes
for symmetric and asymmetric key crypto systems are 80 and 160 bits, respectively.
For symmetric key crypto system, National Institute of Standards and
Technology (NIST) of USA proposed security level of 80-bit key to be phased out by
year 2015 and used until year 2010 (E. Barker, W. Barker, Burr, Polk, & Smid,
2007a, 2007b). US government has an export policy to control the power of
cryptographic algorithm by setting the maximum key size. The current export limit
of symmetric key size has been raised from 40 bits to 128 bits.
For the symmetric key crypto system of Advanced Encryption Standard
(AES), there are three key sizes: 128， 192， and 256 bits. The asymmetric key
crypto systems, which demand for the minimum private key size at 160 bits by year
2010， are unite field cryptography (FFC) and elliptic curve cryptography (ECC).
FFC and ECC are based on the mathematical hard problems of discrete logarithm
problem and elliptic curve discrete logarithm problem, respectively. The
corresponding sizes of private keys to the AES are 256， 384， and 512 bits,
respectively (E. Barker, W. Barker, Burr, Polk, & Smid, 2007a, 2007b; Gehrmann &
Naslund, 2005， 2006， 2007). The symmetric key is normally remembered by brain;
whereas the asymmetric private key is encrypted using another symmetric key.
50
ASCII characters have entropy of 6.57 bits per character. Therefore, the
nominal bit of an ASCII character is 8 bits, but its effective bit is 6.57 bits. To cater
for the different symmetric key sizes at 80， 96， 112， 128， 192， and 256 bits as in
Table 3.3， 13， 15， 18， 20， 30， and 39 ASCII characters are needed, respectively. An
amount of 15 ASCII characters is perhaps still affordable and convenient for the
human users. However, higher amounts may introduce two problems.
Memorizability is the main problem. The difficulty to type a long passphrase into a
computer will be another open problem (Schneier, 1996).
Here, a high-entropy key input method called 2-dimensional (2D) key as in
Figure 4.4 is proposed to solve these problems. 2D key facilitates particularly the
recognition of reference points of each sub-unit of a passphrase like CLPW of CLPP,
and generally the creation of various secret styles of 2D key like multiline passphrase,
crossword, ASCII graphic s/art, Unicode graphic s/art, colorful text, sensitive input
sequence, and two or more of their hybrid combinations as partially illustrated in
Figures 4.6-4.9， for Latin language users.
It uses a 2D display as user interface to improve the human factors of
memorizability and input of ASCII characters from keyboard. The 2D key has the
styles of multiline passphrase, crossword, ASCII art, colorful text, or sensitive input
sequence. It can resist dictionary attack and fulfil the demands of human-
memorizable key sizes even until 256 bits, which is impossible by using the single-
line passphrase. CLPW and CLPP may also be used in the 2D key.
In addition to fulfilling the various key sizes of symmetric key crypto system,
2D key has novel revolution to the private key storage of asymmetric key
crypto system. For the prior arts, we have encrypted private key, split private key, and
roaming private key. With the introduction of 2D key, there shall be no more need to
store the private key in a computing system, but inside the brain as like the
symmetric key. This allows the creation of memorizable public-key crypto system
(MePKC) as discussed in Chapter 9. MePKC has the special features of mobility,
lower cost, and higher efficiency.
51
400
V
Optionally activate the anti-keylogging software.
、
f
Open the 2D key application software for its input interface:
(1) Select row and column sizes
(2) Select to view or hide the secret to be entered
、
f
401
/X7I
402
User enters a secret into the 2D field using one or a combination
of the secret styles as follows:
(1) Multiline passphrase
(2) Crossword
(3) ASCII graphics/art
(4) Unicode graphics/art
(5) Colorful text
(6) Sensitive input sequence
(7) Other hybrid combinations
/X7I
403
Further secret processing over the password using the optional
techniques as follows in sequential order or not in order:
(1) Key hashing
(2) Key strengthening
(3) Multihash key
(4) Other secret processing techniques over the password like
generating multiple slave keys from a master key
、
f
Apply the finally generated secret(s) for various applications.
、
f
Clear the memory storing the initial, intermediate, and final
secrets. Then, close all the application software.
ᄉ \71
404
ᄉ X71
405
406
Figure 4.4 Operation of 2D key input method and system
52
4.3.1 Related Work: Single-Line Key/Password Field
Conventionally, whenever secret is used as the authentication method, single-
line key field will be the area for a user to enter a key. For the current longest
possible key, it is a single-line passphrase. For passphrase, it can be formed from
acronym, sentence, diceware, and coinware. Nevertheless, there is a limit due to the
problems of memorizability and ASCII character input from keyboard. The first
problem is due to the human factor; whereas the second is due to the user interface.
These problems prohibit the applications of symmetric key sizes at higher security
levels whenever a user cannot remember and/or conveniently enter a long single-line
passphrase.
4.3.2 Related Work: Key Strengthening
Key strengthening is also called key stretching. It is used to make a weak key
stronger. There are two forms of key strengthening. One uses password supplement
(Manber 1996; Abadi, Lomas & Needham, 1997; Abadi, Needham & Lomas, 2000)，
and another uses many rounds of hasn iterations (Kelsey, Schneier, Hall & Wagner,
1997). In this thesis, key strengthening is applied to achieve larger protection periods
for symmetric and asymmetric crypto systems like AES and MePKC.
S = n*L*R/P (4.1)
S = Key space
n = Number of networked computers
L = Maximum lifetime of a key in years
R = Number of guesses per unit of time per unit of computer
P = Probability that a key can be guessed in its lifetime
Typical values:
n= 10 9 units = 29.9 bits
53
L = 4， 10， 20， 30， 300 years = 2， 28.2， 29.2， 29.8， 33.1 bits
R = 1.5 X 10 7 s— 1 = 23.8 bits (best performance in year 2005)
R = 1 s _1 = bit (using key strengthening)
P= 10— 6 = -19.9 bits
Equation (4.1) is a password length equation. When key strengthening is used,
R becomes 1 guess per second and the variety of computer is a main factor to set the
number of hash iterations. The computer performance of a variety of computers
varies from 1 time for the slowest computer to 20 times for the fastest computer. This
contributes a factor of log220 = 4.3 bits to Equation (4.1). Moore's Law states that
the number of transistors on an integrated circuit for minimum component cost
doubles every 24 months (Wikipedia Contributors, 2008v).
S = (n*L*R/P)* 2 43 * 2 L/2 (4.2)
When the variety of computers and Moore's Law are considered, it becomes
Equation (4.2). From Equation (2)， key strengthening can make a weak key to
become 19.5 bits stronger.
4.3.3 2D Key Input Method
For single-line passphrase, the numbers of ASCII characters for different
symmetric key sizes are shown in Table 3.3. An amount of 15 ASCII characters or
96 bits is a memorizabilty limit for many human users. This fact is statistically
proven by Florencio and Herley (2007) in their large-scale study of web password
habits for half a million users over a 3-month period, where the average key size is
40.54 bits ranging from exclusive to inclusive 100 bits or ]0， 100]. The difficulty of
user interface to enter a key using keyboard into the single-line key field is another
big problem.
54
The problems of human factor and user interface limit the practical
application of symmetric key crypto system to be at the key size of 96 bits with 10
years of protection. Using key strengthening, the 96-bit key can be made 19.5 bits
stronger, and 20-year protection is the maximum theoretical limit.
The 2-dimensional (2D) key input method is created to allow high-entropy
keys. Figure 4.5 displays the pseudocode of 2D key input method. It tries to solve the
human factor of memorizability and user interface of key input. 2D key has a 2-
dimensional display alike a 2D matrix, where each character of a key is an element
of the matrix.
1.0 User selects row size.
2.0 User selects column size.
3.0 User enters ASCII characters or Unicode symbols one by one.
4.0 User ends the key input by pressing the "Enter" key.
5.0 Computer hashes the input key.
6.0 Computer compares the hashed key with the stored hash.
6.1 If the hashes match, authentication is verified.
6.2 If the hashes mismatch, authentication is rejected.
Figure 4.5 Pseudocode of 2D key input method and system
The font used for 2D key has to be fixed-width font (Wikipedia Contributors,
2007f). Fixed-width font is also called non-proportional font and monospaced font. It
is a typeface using fixed width for every glyph. Examples of fixed-width fonts are
Courier for ASCII and MS Mincho for Unicode. When ASCII encoding is used, the
2D key has 6.57 bits per character. Meanwhile, when Unicode is used, it has 16 bits
per character. Even though Unicode-based 2D key has higher entropy, it is
inconvenient to enter a Unicode symbol for the mean time, and the fixed-width font
for all the Unicode symbols has not yet been created. Hence, ASCII-based fixed-
width font is used currently for the discussions as well as prototype demonstration.
To use 2D key input method and system, firstly a user needs to select the row
size and column size of the 2D matrix for 2D key. The currently built prototype has a
maximum row size or height of 10 characters, and a maximum column size or width
55
of 13 characters. The column size is set at 13 due to the Chinese-character-encoded
passphrase proposed in Section 4.2 has a maximum size of 13 per Chinese character.
Alternatively, it can be a word in English language or other languages that has a size
of 13 characters per word with character stuffing.
The input styles of 2D key are multiline passphrase, crossword, ASCII art,
Unicode art, colorful text, and sensitive input sequence. Multiline passphrase,
crossword, and ASCII art are currently implemented in the prototype; whereas
Unicode art, colorful text, and sensitive input sequence require additional supports.
After selecting the row size and column size, the user can input ASCII
characters using keyboard as the elements of the 2D matrix. The input characters can
have any style or a mixed style of 2D key. These styles have good memorizabilty,
and the 2D nature of 2D key generates more references at the user interface for key
input. Single-line key field has only one reference at the first location of the only line.
2D key has a number of horizontal lines and each first location of the horizontal lines
acts as references for key input. In addition, the first locations of the vertical lines
can be secondary set of references for key input. This solves the location recognition
problem of user interface in facilitating a user to enter a high-entropy key by having
more indexed references.
Good memorizability allows a user to repeat a high-entropy key. The
elements of 2D matrix can be either partially, fully, or extraordinary filled. To fill
extraordinarily means adding some extra trailing characters as noise after the last
element of the 2D matrix. The characters entered into the 2D key field are read by a
computer line by line horizontally from top to bottom, hashed, and processed as
usual alike the single-line key field. The hashing process is one round if key
strengthening is not used. If key strengthening is used, the hashing iteration is set
according to the computer response time per access ranging from 0.05 to 1 second, or
any other tolerable ranges.
The advantages of 2D key are good memorizability, high-entropy key, more
references at the user interface to facilitate key input, and resistance to dictionary
attack. Even pre-computation attack can be avoided if the 2D secret is used on the
platform of MePKC. Its disadvantages are more time for key input and possible
56
shoulder- surfing attack. Nevertheless, for a long passphrase having many individual
units like word, the key input time of 2D key is faster than the single-line key field
whenever there is some interrupt and the user has forgotten the input sequence. This
is because only that particular sub-unit has to be re-keyed in and not the whole secret,
such like the secret style of multiline passphrase.
4.3.4 Styles of 2D Key: Multiline Passphrase
For single-line key field, it is hard to input a high-entropy single-line
passphrase due to the problem of user interface. A user may lose the reference of
starting character of a word in a passphrase. Using 2D key, multiline passphrase can
be input, where each line consists of one word of a passphrase. Each word is padded
to the longest word in the passphrase. The padding character can be any ASCII
character and acts as a text-based semantic noise. Figure 4.6 shows a 2D key
example using multiline passphrase. Its dimensions are 4x5, and uses character
as the padding character. This 2D key has entropy of 131 bits.
Have^
7t 7t 7t
happy
day ！ *
Figure 4.6 Styles of 2D key: Multiline passphrase
HAPPY ᅮ
M^INCH
SPELLS
Figure 4.7 Styles of 2D key: Crossword
4.3.5 Styles of 2D Key: Crossword
The second style of 2D key is crossword. Instead of horizontal and vertical
multiline passphrase, a user can enter a mixture of horizontal, vertical, and slanted
57
passphrases. Figure 4.7 shows two 2D key examples using crossword. Their
dimensions are 5 x 6 (left) and 5x7 (right), and use characters and ':，，
respectively, as the background character. These 2D key have entropy of 197 and
229 bits, respectively.
4.3.6 Styles of 2D Key: ASCII Art / Unicode Art
The third style of 2D key is ASCII art or Unicode art. ASCII art is a graphical
presentation of computer using the 95 printable ASCII characters (Wikipedia
Contributors, 2007g). Unicode is a variant of ASCII art, where instead of using
ASCII characters, Unicode symbols are used to create artistic graphics.
111111
111111
— — 11 — —
— 11 —
— 11 —
111111
111111
Figure 4.8 Styles of 2D key: ASCII art
Figure 4.8 displays three 2D key examples using ASCII art. For the left
example, ASCII characters 'V and are used to display a Chinese character
meaning "engineering". Its dimensions are 7x6 with entropy of 275 bits. For the
middle example, ASCII characters '2' and 'D' are used to display a digit '10' with
background character '2，. Its dimensions are 5x6 with entropy of 197 bits. For the
right example, ASCII characters T and are used to display a Chinese character
meaning "center" with background character Its dimensions are 5x6 with
entropy of 197 bits.
Figure 4.9 shows a 2D key example using Unicode art. Unicode symbols
and are used to display a Chinese character meaning "engineering" again.
Unicode '¥' is entered using the keyboard by pressing the keys "0165" while holding
the key of 'Alt'. Unicode is entered using the keyboard by pressing the keys
222222
2D2DDD
2D2D2D
2D2D2D
2D2DDD
* * * * I
58
"0169" while holding the key of 'Alt'. Once the 'Alt' key is released, the Unicode
symbol is entered. Its dimensions are 4x5. This 2D key has entropy of 320 bits.
¥¥¥¥¥
©©¥©©
©©¥©©
¥¥¥¥¥
Figure 4.9 Styles of 2D key: Unicode art
4.3.7 Styles of 2D Key: Colourful Text
The style of this 2D key needs some additional supports. Color encoding,
special graphical user interface, and special computer processing are required.
Although these supports make the user interface complicated for the computer, they
can be implemented and have better memorizability for the human users. Color is
definitely a main element of good memorizability. For example, by having 16 types
of colors, every character in the 2D key will have an additional 4 bits. ASCII-based
2D key will become 10.57 bits per character; whereas Unicode-based 2D key is
20.59 bits per character. The entropies per character of ASCII-based and Unicode-
based 2D keys will be increased by 60.9% and 24.1%, respectively. The additional
color secret also carries more randomness to resist dictionary attack.
4.3.8 Styles of 2D Key: Sensitive Input Sequence
For the secret style of sensitive input sequence, it is an additional feature over
the current 2D secret style where there is added entropy from the input sequence of a
character to a specific element location of the 2D matrix. If a 2D key has the
dimensions of m x n， the key space is increased by [(m * n)l]. If a 2D key of 4 x 5 as
in Figure 4.4 is used, the key space is increased by [20!] or 61.1 bits from 131.40 bits
to 192.47 bits, which is close to the left example in Figure 4.5 for the 2D key of
dimensions 5x6 with 197.10 bits.
This style requires the space encoding for the element location of 2D matrix,
table- like graphical user interface of m x n matrix, and human memory for the
59
sequence of characters. In term of memorizability, there is not much improvement.
However, the time to enter a 2D key of similar size is greatly reduced for the same
amount of entropy.
4.3.9 Applications for Symmetric and Asymmetric Key Cryptosystems
With the emergence of 2D key having the styles of mutliline passphrase,
crossword, ASCII art / Unicode art, colorful text, and sensitive input sequence, high-
entropy key as high as 256 bits is possible. Chinese-character-encoded passphrase
can be efficiently used for the 2D key style of multiline passphrase. We can now
overcome the human factor of memorizability and user interface problem of single-
line key field, which have limited the key size to 96 bits.
Table 4.16 shows the possible dimensions of ASCII-based 2D key for various
key sizes of symmetric key crypto system. Key strengthening can boost up another
19.5 bits. If Unicode-based 2D key is used, the dimensions of 2D key can be greatly
reduced. From Table 3.3， the settings sufficiency of some key input methods and
systems for various key sizes is shown. It can be observed that larger key sizes than
128 bits for cryptographic, information-hiding, and non-cryptographic applications
like AES-128, AES-192, AES-256, ECC-256, etc., can be realized by using the 2D
key, especially the MePKC using fully memorizable private key.
Table 4.16 Dimensions of 2D key for various symmetric key sizes
Symmetric key size (bits)
80
96
112
128
192
256
Number of ASCII characters
13
15
18
20
30
39
Dimensions of 2D key
3x5
3x5
3x6
4x5
5x6
5x8
For asymmetric key crypto system, memorizable public-key crypto system
(MePKC) can be created. This is possible by using the FFC and ECC with minimum
size of private key at 160 bits. The private key of MePKC is stored in the human
brain, and not stored as encrypted, split, and roaming private keys as in the prior arts.
This provides mobility, lower cost, higher efficiency, and resistance to dictionary and
pre-computation attacks.
60
Assuming that the maximum memorizable key size is 256 bits, 256-bit
MePKC using FFC and ECC with 128-bit security strength can be realized. It has a
protection period of 30 years. If key strengthening is used, 19.5 bits is added, or an
increase of 10-bit security, which extends the protection to 50 years. This is very
much enough for many practical applications. For more information, please refer to
Chapter 9.
A software prototype of this 2D key (Lee, 2006b, 2008i) with the function of
multihash key (Lee, 2007a) has been built up by using the Microsoft Visual Studio
(Marshall, 2003). The 2D key can have optional anti-keylogging application software
(Log This, No date; McNamara, 2003， pp. 197-202) to achieve higher security during
the input. To get a copy of this software, please visit [URL: www.xpreeli.com].
There are other potential applications of 2D key methods and systems. Firstly,
2D key can be specialized to include only numeric digits or other sets of limited
encoded characters for devices with limited space like the display and key pad of a
bank ATM machine and computerized safety box. Secondly, the display of 2D key
can be an LCD display or other display technologies integrated with a computer
keyboard having a first partial 2D key optionally visible and a second partial ID key
in hidden mode only to better resist the shoulder- surfing attacks.
4.3.10 Conclusion
Here, the high-entropy 2D key input method has been proposed. It solves the
memorizability problem due to human factor and user interface problem of single-
line key/password field. Chinese-character-encoded passphrase is efficient for the 2D
key style of multiline passphrase. Besides, 2D key has the styles of crossword,
ASCII art, Unicode art, colourful text, and sensitive input sequence. The
memorizable limit of 96-bit key is increased to 256-bit key, where even the private
key is memorizable. This creates 160-bit to 256-bit MePKC with protection period
up to 50 years.
61
CHAPTER 5 CREATING BIG MEMORIZABLE SECRET
(PART 3)
5.1 Multilingual Key
2D key is a method suitable for Latin language users. For users of CJK
languages using the Han character and other non-Latin languages, multilingual key is
an alternative to 2D key. For the trends, statistics, and geo-political coverages of
monolingual, bilingual, and multilingual language users, please refer to Section 4.1.5.
5.1.1 Related Works
For the related works of multilingual key, please refer to Section 3.4 on
potential methods to create big and yet memorizable secret. These related works are
graphical password, Passfaces, Draw-a-Secret scheme, icon-like graphical password
scheme, and event-based graphical password scheme.
5.1.2 Black-and-White Multilingual Key
For multilingual key, graphical password/key method and system is somehow
innovated to have both the features of cognometrics and locimetrics by using graphic
symbols of multilingual languages from any symbol encoding code, such as Unicode,
specifically. This invention is especially effective for logographic, bilingual, and
multilingual language users. In this new secret creation method, there is a huge key
space comprising black-and-white and/or colorful Unicode graphic symbols grouped
into tabular pages as in Figure 5.1 illustrating one of the exemplary tabular pages
{4E00-4EFF}.
For this black-and-white multilingual key, a user knowing a particular
language has the property of cognometrics to recognize a graphic symbol.
Furthermore, there exists also the property of locimetrics for a user to locate a tabular
page, subsequently a graphic symbol, and finally a partitioned area of a Unicode
graphic symbol.
62
500
4E
o
1
2
3
4
6
7
8
9
ノ
A
B
c
D
E
F
丐
丰
乐
习
买
ノ 、
亀
京
ノ j 、
什
1 1
牛
仔
1
丁
-*
丑
両
\
、
舌し
1 ― 1 ᄂ i
乾
丁"
亡
{H
1 •■
仁
ィ乞
ィ今
2
丢
串
1
乂
丘
亂
互
7 ᄂ
ノ u
쟈 T.
ノ 1、
仿
しん J
令
、
fg
仲
1 1
3
七
专
ノ 1
弊
ノ J
乃
丘
ᄀᄂ i
乳
-J ᄂ i
舞 ᄂ
/T ᄂ i
fr
ノ 1
7T
ノ j
毫
1 ~i
仓
代
1 Ni
4
丄
且
两
1 ~}
临
1 M-4
メ
豊 ᄂ
五
父
JX
！ノ V
仔
1 -J
f 月
5
조
严
半
1
ズ
辛し
-J
井
ノ 1
褒
"도 ᄂ、
仅
し/ V
仕
1_ᄂ
以
""ノ 、
仔
1 1
6
世
1 ——
並
、
9、
5i 匕
书
■ ~ノ
r L
了
亦
直
仆
1 1
他
f 小
1 -J
件
1 1
7
万
f 什
1— H
丧
レ、
、 /
튄
乗
ノ 1 、
C i
' L
产
廉
IA 1、
仇
仗
价
1 ノ 1
8
丈
ノ^^
丘
丸
ノ 1— i
乘
^ 1 、
台 ᄂ
1 ~ 1 ᄂ i
碰
予
豆
t __ >
|$
つ 1
付
1 ᄀ J
伏
\y 、
9
丙
1\
1
丹
乙
占 ᄂ
1 ᅳ 1 u
草 ᄂ
ᅳ
争
瓦
亩
1 ■ 1
仙
让
伴
1 1
A
上
业
个
— -
、 /
ᄂ
m
事
亚
N.
人
y 、
I
1^
1 1 1
合
B
下
丛
丫
1
主
— y
カロ
事
ifcb
i
I
介
ノ 1
1 1— i
任
1 ~ ᄂ
C
ᅲ
东
井
乌
"a
亜
京
fx
任
D
丝
中
乍
九
右
ムみ
个
仍
fit
份
E
丞
举
乞
乾
亞
[A
m
仮
但
F
丐
丢
丰
J
乏
也
s ᄂ
亿
仏
if
抄
仿
Figure 0.1 une of the exemplary tabular pages of multilingual kev consisting of tne
first 256 Han characters in the Unicode and starting from Unicode value {4E00}
63
The input method of multilingual key is normally a computer mouse, where it
can also be other input devices like touch screen, tablet, stylus, keyboard, sound
recognition, eye-tracking technology, Microsoft Surface, etc. The monitor tend
towards wide- screen LCD at lower cost shall popularize the multilingual key.
5.1.3 Grid Partitioning for Higher Entropy and Randomness
To increase the entropy per image selection and its randomness to resist
guessing attack and dictionary attack, invisible grid partitioning is applied to every
graphic symbol based on the setting of 3 ^ 3, particularly, or any other settings such
as 2 ^ 2, 4 ^ 4, and so on, as in Figure 5.2.
Figure 5.2a Figure 5.2b
Figure 5.2c Figure 5. 2d
Figure 5.2 A Han character from Unicode before and after the grid partitioning for
various settings: (Figure 5.2a) Without grid partitioning; (Figure 5.2b) with grid
partitioning of 2 ^ 2; (Figure 5.2c) with grid partitioning of 3 ^ 3; and (Figure 5. 2d)
with grid partitioning of 4 * 4
3
o
64
These partitioned areas increases the entropy of multilingual key by 2, 3, and
4 bits, respectively, for 2 * 2, 3 * 3， and 4*4 settings. Every partitioned area
represents the concatenation of a few bits to the bitstream encoding a graphic symbol
using Unicode in a tabular page consisting of 256 symbols or flexibly any other
amount. Among the settings of grid partitioning, 3 * 3 is selected as the optimum
settings and used for further explanation.
1
2
3
blank
4
6
7
702
8
9
10
11
blank
12
13
14
15
ᄉ X7I
703
Figure 5.3c
Figure 5.3d
Figure 5.3 und partitioning encoding of a graphic symbol, wherein (Figure 5.3a)
illustrates the 3*3 settings where red lines are invisible; (Figure 5.3b) illustrates the
encoding for human memorization and reference in the human context; (Figure 5.3c)
BMP (Basic Multilingual Plane) when a partitioned area is selected in the computer
context; and (Figure 3.jd) illustrates the concatenated bit values to the Unicode value
of a graphic symbol in the SIP (Supplementary Ideographic Plane) when a
partitioned area is selected in the computer context
65
There are nine partitioned areas in the setting of 3 * 3 as in Figure 5.3a. The
outer 8 partitioned areas are encoded by 3 bits. Meanwhile, the central partitioned
area adds no bit. For Han characters and other multilingual languages, two Unicode
planes are used in the multilingual key, where more Unicode planes can also be
added. These are BMP (Basic Multilingual Plane) and SIP (Supplementary
Ideographic Plane), where both can support 65536 (= 216) graphic symbols.
For computer context, graphic symbols from different Unicode planes are
encoded by bit for BMP and bit 1 for SIP; whereas the 9 partitioned areas have the
central area to carry blank value, and the outer areas to represent bit values of 0， 1， 2，
to 7 for BMP and 8， 9， 10， to 15 for SIP, as in Figures 5.3c and 5.3d, respectively.
For human context, to ease memorization and references, the 3*3 partitioned areas
are again encoded by digits from 0， 1， 2， to 9 as in Figure 5.3b. The central area
represents digits and 5; whereas the outer areas represent 1, 2, 3, 4， 6， 7， 8， and 9
for both graphic symbols from BMP and SIP. Hence, the 3 x 3 grid partitioning adds
either bit with one-fifth (1/5) probability, or 4 bits with four-fifth (4/5) probability,
to the Unicode value of a selected graphic symbol.
For instance, for a Chinese language secret of [秦 漢] (Qin Han), the code of
multilingual key without grid partitioning is {79E66F22}i 6 based on Unicode, where
{79E6}i6 represents [秦] (Qin) and {6F22} 16 represents [漢] (Han). When 3 * 3 grid
partitioning is used, two more digits of secret are added. Let the first digit to be {4}i
to represent the western piece of partitioned areas of [秦] (Qin), and the second digit
to be {5} io to represent the central piece of partitioned areas of [漢] (Han).
Consequently, the constructed secret is [秦 4 漢 5] (Qin 4， Han 5).
Since both the Han characters of [秦 漢] (Qin Han) are in the BMP, then the
encoded secret for a computing device is {79E636F22}i6. The concatenated
hexadecimal digit of {3}i6 to the end of the Unicode value of {79E6}i6 is constructed
from {001 1 )2 where the first bit represents the BMP and the last three bits represent
the western piece of partitioned areas. For the numeric secret of {5}io, no
hexadecimal digit is added because digits {0}io and {5}io represent no concatenated
66
value to the Unicode value of selected graphic symbol. The concatenation of these
numeric secrets representing different partitioned areas can be at any location of the
Unicode values of the selected graphic symbols.
Therefore, for black-and-white multilingual key with 3*3 grid partitioning, a
selected image by clicking a partitioned area carries 16.59 or 20.59 bits, with
probabilities of 1/5 and 4/5， respectively. For a sequence of many selected
partitioned image areas, the average entropy per image selection for this type of
multilingual key is 19.79 bits.
5.1.4 Colourful Multilingual Key
To further increase the key space for higher entropy, colourful multilingual
key is an added option. The (16+l)-colour scheme of colourful multilingual key as in
Figure 5.4 is selected for explanation, where it can also be other settings. The (2+1)-，
(4+1)-， (8+1)-， and (16+l)-colour schemes of colourful multilingual key additionally
add 2， 4， 6， and 8 bits, respectively, to the black-and-white multilingual key with 3 *
3 grid partitioning. This means that a selected partitioned image area of (16+1)-
colour multilingual key has 24.59 or 28.59 bits and average entropy of 27.79 bits.
Also, besides Unicode character and partitioning digit, a user needs to remember a
third secret for the combination of foreground and background colours.
Yet to further increase the key space, some special text processing techniques
can be used, wherein examples include special effects like directional shadow, 3D
styles, and lighting; enclosed character using shapes like circle, square, triangular, or
diamond; typeface variation like font type, font size, as well as font format of single
strike through, double strike through, and underscore/underline; mirror images on the
left, right, up/down; 45°-, 90°-, and 135°-degree clockwise and anti-clockwise rotated
images; solid and hollow images; and background watermark.
5.1.5 Font Technologies to Solve Data Size Problem
Nevertheless, the potential huge key space of colourful multilingual key with
and without special text processing techniques has memory storage problem due to
67
its huge image size if tabular pages of graphic symbols are stored in normal image
file format like BMP, GIF, JPG, and PNG. For black-and-white multilingual key, its
problem is not the image storage, but the image loading to the limited RAM, which is
also a second problem to the colourful multilingual key.
800
p
f
星
星
星
星
e
星
星
星
星
星
n
d
星
星
星
星
r — »
m
c
星
星
星
星
1
b
星
星
星
星
k
a
星
星
星
星
星
J
9
星
星
星
星
星
星
も
星
星
星
星
星 星
ᄂ
n
7
星
星
星
星
g
6
星
星
星
星
星
星
星
星
星
星
星
星
星
星
星
f
5
星
星
星
星
e
4
星
星
星
星
d
3
星
星
星
星
c
m
星
星
n
星
星
星
星
星
b
星
m
星
星
星
星
星
a
星
星
星
星
星
星
星
星
星
星
星
星
星
星
1
2
3
4
5
6
7
8
9
a
b
C
d
e
f
a
b
c
d
e
f
g
h
j
k
1
m
n
p
Using XI i Color names or any other color encoding:
Black #000000 Brown #A52A2A Red #FF0000
Yellow #FFFF00 Green #008000 Blue #0000FF
Gray #808080 White #FFFFFF Silver #C0C0C0
Salmon #FA8072 Gold #FFD700 Khaki #F0E68C
Pink #FFC0CB
Orange #FFA500
Violet #EE82EE
Tan #D2B48C
Cyan #00FFFF
N.B. (Nota Bene): The first 10 encoded colours from to 9 are according to the resistor colour
code. Other extended digits from 10 to 15 are the lighter colours from black to green, respectively
modulus 10, and the last colour pink is used as the front-slash- wise diagonal background colour.
Figure 5.4 ᄂ lb+l)-colour scheme for colourful multilingual key
68
To solve the limited RAM problem of black-and-white and colourful
multilingual keys, the image file format of PNG (Portable Network Graphics), which
is good for image compression of line art, can be used for efficient size of image
database. Yet for better file compression, algorithm of DJVU file format can be
further applied by splitting a tabular page into many layers for separate compression.
However, the best current possible and practical solution to both the problems is to
have real-time font rasterization from font files like outline font or vector font storing
all the Unicode graphic symbols to the monitor display.
5.1.6 Resistance Techniques to Shoulder-Surfing Attack
Another problem of multilingual key is shoulder- surfing attack from a person
or camera nearby the monitor and able to watch and record the image area selection
of sequential Unicode graphic symbols. The first solution relies on the human
memorizability limit and asks a user to do false selection of image areas by toggling
a key on the keyboard, or single-double or left-middle-right clicking of mouse. The
second solution is to allow a user to enter a textual password into a key field at any
interim session during the input of a graphical key. In other words, the second
solution is a hybrid method combining the textual and graphical passwords/keys.
5.1.7 Fast Search for a Unicode Graphic Symbol
Yet another problem of multilingual key is its huge key space causes the
search of a graphic symbol to be slow if only images of Unicode graphic symbols are
stored. To solve this problem, there can be some tabular pages specially designed to
list and show the frequently used Unicode graphic symbols, especially Latin and Han
characters, or Latin and other languages, to speed up the image area selection of a
Unicode graphic symbol. A second solution is to have a fast input method and
system of Unicode graphic symbol to search and locate the tabular page and specific
location of a particular graphic symbol, which is now possible for Latin languages
and CJKV languages using Han characters.
69
900
V
Optionally activate the anti-keylogging software.
、
Open the multilingual key application software for its input
interface.
、
f
ᄉ X7I
901
902
User enters a secret of multilingual key:
(1) Search for the specific tabular page containing the Unicode
graphic symbol
(2) Real-time font rasterization displays the specific tabular page
containing the Unicode graphic symbol
(3) User optionally clicks on a selected Unicode graphic symbol
to access the (16+l)-color scheme
(4) User clicks on the partitioned area based on digit secret and
optional color secret
(5) User optionally cancels for false signal to resist shoulder-
surfing attack or confirms on the selected secret of Unicode
graphic symbol together with its secrets of digit and color
(6) Repeat steps (1) to (6) of Box 903 in sequential order until
sufficient key entropy has been achieved
/ATI
903
、
User optionally enters another textual password/key into a
password/key space to resist shoulder-surfing attack.
f
Undergo secret processing technique(s) as in Box 404. Then,
apply the finally generated secret(s) for various applications.
/
Clear the memory storing the initial, intermediate, and final
secrets. Then, close all the application software.
904
905
906
Figure 5.5 Operation of multilingual key input method and system
70
5.1.8 Applications for Symmetric and Asymmetric Key Cryptosystems
Subsequently, big memorizable secret for cryptographic, information-hiding,
and non-cryptographic applications in information engineering can be created from
multilingual key as in Figure 5.5 according to the specific demand thresholds for
various key sizes in Table 3.3. More importantly, MePKC using fully memorizable
private key can be specifically realized. For more information, please refer to
Chapter 9.
Multilingual key input can have optional anti-keylogging application
software (Log This, No date; McNamara, 2003) to achieve higher security. The input
method of multilingual key is normally a mouse, where it can also be other input
devices like touch screen, tablet, stylus, keyboard, sound recognition, eye-tracking
technology, Microsoft Surface, etc. The key space is increased using pictorial
colourful Unicode graphic symbols with 17 background colours and 16 foreground
colours, which can also be increased using special effects like directional shadow,
3D styles, lighting, enclosed character using shapes like circle, square, triangular, or
diamond, as well as typeface variation like font type, font size, and font format.
The (16+1) colours of colourful multilingual key in Figure 5.4 are black,
brown, red, orange, yellow, green, blue, violet, gray, white, silver, tan, salmon, gold,
khaki, and cyan for 16 foreground colours, and black, brown, red, orange, yellow,
green, blue, violet, gray, white, silver, tan, salmon, gold, khaki, cyan, and pink for 17
background colours. The first 10 colours of the (16+l)-colour scheme has good
memorizability based on the colour code of resistor. The next 6 colours are lighter
colours than the corresponding colours modulus 10. The last colour pink is used as
the front-slash-wise diagonal background colour. Other colour combinations may
also be possible.
A software prototype of this multilingual key (Lee & Tan, 2006a) has been
built up by using the Microsoft Visual Studio (Marshall, 2003). To get a copy of this
software, please email the author at [Email: E9t3iJCW@hotmail.com]. Whenever the
conditions like time, money, and online archival service allow, the author shall mail
you a CD copy or upload the software.
71
5.2 Multi-Tier Geo-Image Key
To create big memorizable secret, a second new type of graphical
password/key is invented using a hybrid combination of recognition-based
cognometrics and locimetrics over a map, as well as recall-based textual
password/key of a space name and characteristics. This space map can be continents
of Earth, seafloor of oceans, constellations of star sky, and so on.
Let's take the Earth map of continents as an example for multi-tier geo-image
key. The current best GPS (Global Positioning System) resolution for civilian usages
is about 15 meters (m) per pixel. For the latest Google super- satellite GeoEye- 1
(Chen, 2008; Wikipedia Contributor, 2008ay) launched on 6 September 2008， it
captured the first image at a resolution of 41 cm on 7 October 2008 (EDT). For spy
satellite (Wikipedia Contributors, 2008ax)， the best resolution may be less than 2 cm.
Let's take 15 meters (m) per pixel in the calculation. The radius of Earth
globe is r = 6.37 x 10 6 m and its surface area is SEanh = 47ir 2 = 5.099 x 10 14 m 2 .
Assume only 2— 7 of Earth surface is memorizable populated areas like metropolis,
city, town, village, etc. Assume also a pixel represents an area of 15 2 m 2 , and a
partitioned area of Earth map at the first tier has 20 * 20 pixels. At a monitor image
resolution of 800 * 600 pixels, there are 1200 partitioned areas at the first tier of
Earth map. Simple estimation will show that four to five tiers of map are needed to
locate a specific location on the Earth surface after subsequently selected image areas.
Through some calculation, the whole Earth surface including continents and
oceans has a surface area per pixel of S P i xe i = 47ir 2 1 15 2 = 2.266 x 10 12 m 2 /pixel, or
entropy of Eearth = 41.04 bits. Considering a click area of 20 x 20 pixels after image
partitioning, the surface area per click area is Sciick = 4ur 2 1 (15 2 X 20 2 ) = 5.665 x 10 9
m 2 /click area, or entropy of 32.40 bits. When the factor of easily memorizable Earth
space like populated areas is included, the usable Earth surface to create a big
memorizable secret is S mem orizabie = 2" 7 x S c ii C k = 4.426 x 10 7 m 2 /click area, or entropy
of 25.40 bits. Hence, a partial image secret of multi-tier geo-image key has about
25.40 bits.
In addition to a partial image secret of a space, a user is also required to enter
a second partial textual secret related to the name and/or characteristics of that
72
particular selected image space or location. This is used to increase the key entropy
and to resist the shoulder- surfing attack. For every partial image secret, there shall be
a partial textual secret. Preferably, the key length of the partial textual secret is at
least 6 characters.
1000
V
Optionally activate the anti-keylogging software.
1001
Open the application software of multi-tier geo-image key for its
input interface showing an Earth map, ocean seafloor, or others.
1002
User enters a partial image secret of multi-tier geo-image key:
(1) Beginning with a first tier of Earth map showing all the continents with
resolution 800 * 600 pixels, select a first partitioned area of about 20 * 20
pixels, for a second tier of map, or as a secret and go to Box 1004 directly
(2) From a second tier of Earth map, select a second partitioned area of about 20 11
20 pixels, for a second tier of map, or as a secret and go to Box 1004 directly
(3) From a third tier of Earth map, select a third partitioned area of about 20 * 20
pixels, for a third tier of map, or as a secret and go to Box 1004 directly
(4) From a fourth tier of Earth map, select a fourth partitioned area of about 20 *
20 pixels as a secret and go to Box 1004 directly
1003
\レ
Yes
User enters a textual password/key related to the selected area
for higher entropy and resistance to shoulder-surfing attack.
f
If the key entropy is still insufficient, go to Box 1003 again and
select another geo-image area ana its related textual key.
No
f
Undergo secret processing technique(s) as in Box 404. Then,
apply the finally generated secret(s) for various applications.
/X ᄀ
1004
1005
1006
\レ
Clear the memory storing the initial, intermediate, and final
secrets. Then, close all the application software.
Figure 5.6 Operation of multi-tier geo-image key input method and system
세 007
73
If ASCII encoding is used, then the textual password/key adds another 39.42
bits. In total, a unit of multi-tier geo-image key has entropy of 64.82 bits. Some units
of multi-tier geo-image key are sufficient for many applications using secret. To
specifically realize the MePKC, three and four units of multi-tier geo-image key can
support 160- and 256-bit MePKC, respectively, using the ECC. The monitor tend
towards wide- screen LCD at lower cost shall popularize the multi-tier geo-image key
as well.
Table 3.3 shows the required unit of geo-image key for various key sizes, and
Figure 5.6 illustrates the operation of this method. The space map can optionally
have invisible and/or visible grid lines for easy references. The input method is
normally a mouse, where it can also be other input devices like touch screen, stylus,
keyboard, sound recognition, eye-tracking technology, Microsoft Surface, etc. To
further increase the key space of this method, the preceding tiers of geo-image key
before the last tier can be included, and early secret selection of larger geographical
area is allowed. Multi-tier geo-image key input can also have optional anti-
keylogging (Log This, No date) application software to achieve higher security.
To further increase the key space of this method, the preceding tiers of geo-
image key before the last tier can be included, and early secret selection of larger
geographical area is allowed. Yet another method to increase the key space is to
invest more resources to recruit the architects to draw the geographical map of
populated areas using the architectural normal scaling of 1:500 (or 1 cm : 500 cm, or
1 cm : 5 m)， which is a resolution better than the civilian GPS resolution 15 m/pixel.
5.3 Multi-Factor Multimedia Key Using Software Token
To create big memorizable secret, especially for MePKC realization in
Chapter 9， the key sizes larger than 256 bits, such like 384 and 512 bits, are hard to
be memorizable, and a possible solution is multi-factor multimedia key using
software token as in Figures 5.7-5.8. For instance, 512-bit MePKC using ECC is
needed to realize the bits of security at 2t>b bits and to resist future quantum
computer attack. Hence, multi-factor multimedia key using software token is
74
invented to halve the memorizable key sizes at equivalent security levels, especially
designed for MePKC operating on the FFC or ECC.
1100
Optionally activate the anti-keylogging software.
Open the application software of multi-factor key using software
token for its input interface.
1101
1102
V
User creates an n-bit secret S like 256 bits using one or more methods as follows:
(1) Self-created signature-like Han character for CLPW and later CLPP
(2) ASCII-based 2D key
(3) Unicode-based 2D key
(4) Multilingual key
(5) Multi-tier geo-image key
(6) Conventional secret creation methods and other future methods
ᄉ X71
1103
User creates a software token T by following the steps as below:
(1) User creates and/or compresses a big electronic multimedia data file, be it
random or non-random bitstream, text, image, audio, animation, video, or
hybrid combinations
(2) User hashes the processed data file using 2n-bit hash function like SHA-512
(3) User encrypts the hash value H of multimedia data ule, using n-bit secret like
256 bits and n-bit AES like AES-256, to create the software token T
(4) To use the multi-factor key K MF ， decrypt T using memorizable secret S to
retrieve hash value H, and hash the concatenation of S and H to produce K MF
K MF ^Hash(SIIH)
1104
V
User stores the software token locally in a storage device like
USB flash anve or remotely in a server for roaming purposes.
1105
Clear the memory storing all forms of secrets. Delete or hide the multimedia data
file and its processed data file. Then, close all the application software.
1106
Figure 5.7 Software token generation of multi-factor multimedia key input method
and system
75
1200
\k
Optionally activate the anti-keylogging software.
N レ
Open the application software of multi-factor key using software /^\J7I
token for its input interface.
1201
1202
\レ
User creates an n-bit secret S uke 25b bits using one or more methods as follows:
(1) Self-created signature-like Han character for CLPW and later CLPP
(2) ASCII-based 2D key
(3) Unicode-based 2D key
(4) Multilingual key
(5) Multi-tier geo-image key
(6) Conventional secret creation methods and other future methods
1203
V
User uses a software token T by following the steps as below:
(1) If the software token is in a local storage device like USB flash drive, a user
loads the software token from the storage device
(2) If the software token is in a remote server, a user downloads the software
token through roaming network
(3) User decrypts the software token T using n-bit secret S to get hash value H
(4) Hash value H optionally undergoes secret processing technique(s) together
with S as in Boxes 404 to become 2n-bit multi-factor key K MF
K MF ^Hash(SIIH)
1204
Apply the finally generated secret(s) of 2n-bit multi-factor key K
for various applications.
1205
\レ
Clear the memory storing all forms of secrets. Then, close all the
application software.
1206
Figure 5.8 Software token acquisition and application of mult i- factor multimedia key
input method and system
76
5.3.1 Related Work
A closest related work to multi-factor multimedia key has been published and
found after the patent filing of multi-factor multimedia key. This related work is
multimedia password (Mannan & van Oorschot, 2008) using hash value of a digital
object as password.
5.3.2 Using Key the Secret and USB Flash Drive
For 2n-bit ECC, where 2n can be as big as 512， its 2n-bit private key can be
derived from a memorizable secret and a 2n-bit hash value. This 2n-bit hash value is
obtained from the hashing of a big multimedia data file with its size at least 512 bits
by 2n-bit hash function like SHA-512. This multimedia data file may be random or
non-random bitstream, text, image, audio, animation, video, or hybrid combinations.
The 2n-bit hash value is encrypted by an n-bit memorizable symmetric key using li-
bit AES like AES-256 to create a software token. Here, 2n-bit ECC and n-bit AES
have equivalent bits of security strength at n bits in the scale of symmetric key
crypto system. This software token is then stored in a local storage device like USB
flash drive, floppy disk, CD-ROM, DVD, etc., or in a remote server.
5.3.3 MePKC Application with Optional Split Key Method
Whenever a user needs to use the 2n-bit MePKC like 2n-bit ECC, one is
either to get the local device storing the software token or to download it from a
server through roaming network. Then, by using n-bit memorizable symmetric key S，
one decrypts the 2n-bit software token to get 2n-bit hash value, which is later used
together with S to derive the 2n-bit private key of 2n-bit MePKC. Hence, this bi-
factor key using an n-bit symmetric key and 2n-bit software token can halve the key
sizes of MePKC by sacrificing some mobility.
This method can be extended to become mult i- factor key easily by
undergoing the similar processes in split private key cryptography (Ganesan, 1996b).
For instance, the software token may require bi-factor or multi-factor authentication,
including at least a biometrics factor to access the software token. Also, during the
77
input, multi-factor multimedia key using software token can have optional anti-
keylogging application software (Log This, No date) to achieve higher security.
5.3.4 Software Prototype
A software prototype of this multi-factor multimedia key using software
token (Lee, 2008k) with the function of multihash key (Lee, 2007a) has been built up
by using the Microsoft Visual Studio (Marshall, 2003). The multimedia key input has
optional anti-keylogging application software (Log This, No date) to achieve higher
security. To get a copy of this software, please visit [URL: www.xpreeli.com].
5.4 Hybrid Combinations
So far, there are five methods discussed independently and slightly linked on
how to create big and yet memorizable secret. These five methods are (i) Chinese-
character-encoded passphrase, (ii) 2-dimensional (2D) key, (iii) multilingual key, (iv)
multi-tier geo-image key, and (v) multi-factor multimedia key using software token.
Here, it is to note that these five methods can be applied in hybrid mode together
with the optional keys of password(s) and passphrase(s).
From psychological studies (Standing, Conezio & Haber, 1970; Standing,
1973; Jansen, Gavrila, Korolev, Ayers & Swanstrom, 2003; de Angeli, Coventry,
Johnson & Renaud, 2005)， graphical secret key has better memorizability. However,
there are special attacks on graphical password/key (Thorpe & van Oorschot, 2007).
Nevertheless, there is a human memorizability limit on the number of unique
secrets that a human can have. Adams and Sasse (1999) reported a limit of 4 to 5
keys that were unrelated and regularly used. Meanwhile, Florencio and Herley (2007)
reported a limit at an average of 6.5 textual keys without any relevancy condition on
the selection of a key, except that the set of keys of a user has to be different.
Coming to here, there is a problem on how to increase the number of unique
secrets from a few master keys to some slave keys for both the offline and online
accounts. Please refer to Chapter 7 for a proposed solution called multihash key.
78
CHAPTER 6 MULTIMEDIA NOISES FOR MORE RANDOM
SECRET
6.1 Introduction
This chapter talks on the various types of multimedia noises or errors, which
can be used to create more random secret. Higher randomness means higher
complexity and harder secret cracking. Then, an enhanced frequency analysis to
decrypt the English text is presented. It is followed by the discussion of information
rates and unicity distances for some big secret creation methods. Lastly, a proof is
shown on why secret with noise(s) has higher information rate and unicity distance.
6.2 Semantic Error Occurrences in the Multimedia Communications
Since the introduction of communications technologies from analogue
communications to digital communications, the communication networks have
evolved into the ubiquitous network of Internet that triggers the Information Era. The
data transmission covers all types of media from bit stream, text, image, audio, to
video. To increase the accuracy rate of the communications, the processes of error
detection and error correction are carried out over the received data, which is
majority-wise in the binary form for nowadays communications.
Nevertheless, in this chapter, the second error level occurring in the
multimedia communications is discussed. The first error level is the bit stream;
whereas the second error level is an advanced error called as semantic error (Lee,
2005a). The semantic error occurrence is explained corresponding to the type of
media. Considering the semantic error occurrence, a protection layer can be obtained
in mining the information from the multimedia communications. It is not only
applicable to the digital media in the Silicon World, but the daily physical media as
well. Hence, we can avoid the critical errors in addition to the higher accuracy rate
during the intake process of information. It is highly useful in the field of science,
commerce, history, archaeology, etc. Specifically here, the semantic errors are
applied to add more randomness to the big secret.
79
6.2.1 Introduction
Mass, energy, and information are the three basic elements of the universe.
Meanwhile, the communications of information in the human society is one of the
two main components in the process of human civilization from Stone Age to the
Modern Age. These two components are the languages to communicate the
information and the tools to generate the functions. In the nineteen century, the first
electronic communications technology of telegraph was initialized in year 1837
(Tomasi, 1998; IEEE Communications Society, 2002). Then the analog
communications evolve to the digital communications. A good example of the
ubiquitous network will be the Internet that triggers the Information Era.
Nowadays, the data communications covers all types of media from bit
stream, text, image, audio, to video. The majority of the received data is in the binary
form via the computing devices. Both of the processes of error detection and error
correction are carried out over the received data to improve the accuracy rate of the
communications. Both of these processes have now been very well improved to
handle the first error level of the multimedia errors (Wu, Cheng & Xiong, 2001; Lu,
2002)， which is the bit stream.
The works on the first error level of bit stream have been quite abundant and
very successful for the data transmission in the secure communication systems
(Schneier, 1996, 2000; Stallings, 2006b, 2007) via the applications of many types of
codes and protocols.
Nevertheless, the focus of this section is to discuss on the second error level
occurring in the multimedia communications. This second error level is an advanced
error called semantic error. The capability to detect semantic error will be somehow
able to differentiate among the neutral, true, and false statements. The semantic error
occurrence is explained corresponding to various types of media. With the
consideration of semantic error occurrence, a protection layer is obtained to mine the
information from the multimedia communications.
It is not only applicable to the digital media in the Silicon World, but the
daily physical media as well. Hence, critical errors can be avoided. In addition,
80
higher accuracy rate during the intake process of information can be achieved. It is
highly useful in the fields of science, commerce, history, archaeology, etc. By mining
the critical information of the essential constraints, various models in these fields can
be set up by using the practical intelligence. Specifically here, the semantic errors are
applied to add more randomness to the big secret.
6.2.2 Multimedia Type of Bit Stream
For the multimedia type of bit stream, it is not common in the human society,
except the I Ching (aka Yi Jing) (易 经) ， until the advent of computer technologies
in the 20 th century. There are basically three groups of them in categorizing the
digital binary bit stream (B.1-B.3):
(B.l) Self-executable execution file
(B.2) No n- self-executable execution file
(B.3) Non-executable data file
The examples of (B.l) and (B.2) are the installation and execution files like
self-extractable and non-self-extractable WinZip files. For (B.3), the example is like
the Microsoft Office Word file with the file extension of .doc.
The advancement of error detection codes, error correction codes, and
cryptology (Schneier, 1996, 2000; Stallings, 2006b, 2007) have allowed the full
recovery of binary bit stream without considering the intrinsic information of the
transferred data via the secure communication systems. Therefore, the consideration
of first error level is normally sufficient for this type of multimedia. The demand for
second error level of semantic error for advanced error checking is normally not an
essential issue.
6.2.3 Multimedia Type of Text
For the multimedia type of text, it can be of monolingual and multilingual
articles. Anyway, one language system represents one knowledge system or
81
hyperspace knowledge library. Hence to learn and understand a civilization and
culture comprehensively, the text version has to be learnt in the original language.
For translated version into other language, the factors of polysemy and multimedia
paradox will make the original version and translated version to have various
disputable issues in semantics. Further categorization has two main groups (T.1-T.2):
(T.l)Hard copy
(T.2) Digital copy
Hard copy may be of handwriting and printed copies. Meanwhile the digital
copy is due to the advent of computer technologies. Nevertheless, first error level
checking is not enough for the both of them, especially for the text group of hard
copy. As an information, first error level checking here means to detect and to
correct the text in the style of one word after another word by treating the word as an
independent symbol as for the binary bit stream for its block data. Nevertheless, it is
not enough due to the associative factors of languages (Crystal, 1999; Finegan, 2004)
and linguistics (Matthews, 1997) when the context of text is taken into consideration.
Subsequently, it is essential for the second error level checking to be
implemented on the multimedia type of text. Considering the possible semantic
errors of text (T.a-T.j) as listed below, it is very important to include them in mining
the information from the multimedia type of text.
(T.a) Readability due to handwriting and printing qualities
(T.b) Copying or duplication process
(T.c) Translation process: Phonetics-, symbolics-, semantics-based
(T.d) Punctuation marks
(T.e) Erosion and dirt that change parts of the articles
(T.f) Book typesetting
(T.g) Book binding
(T.h) Usage of ancient, intermediate, and modern language
82
(T.i) Missing supporting references
(T.j) Polysemy where a word has two or more meanings
The significant examples of readability due to handwriting (T.a) are as
follows:
(T.a.l)"DaoDeJing，， (道 德经) (Book of Ethics) and "Yi Jing" (易 经)
(Book of Changes) in ancient Chinese language writing that is being
transformed into the modern Chinese language writing
The significant examples of translation process (T.c) are as follows:
(T.c.l) Books of "Greek Mythology", "Homer's Epics", "Al-Quran" (Tareq
Rajab Museum, 1998-2002), and "DunHuang Story" in the oral form
of human brain memory and human sound
(T.c.2) Books of "The Holy Bible: New Testament" ("The Holy Bible," 1997)
from English language, "The Book of God: Old Testament"
(Wangerin, 1999a) from Hebrew language, and "The Book of God:
New Testament" (Wangerin, 1999b) from Greek language
(T.c.3) Book of "Diamond Sutra" in the translated version of Chinese
language from the Sanskrit language
(T.c.4) Books of "Rig Veda", "Puranas", "Upanishad" (Xu， 1984),
"Mahabharata" (Buck, 2000a), and "Ramayana" (Buck, 2000b) in the
English and Chinese languages translated from the Sanskrit language
(T.c.5) Books of "Arabian Nights", "Kama Sutra", and "The Perfumed
Garden" in the translated version of English language from the
Arabian language
A good example is given as follows to show the effect of punctuation mark
(T.d) via a series of sentences:
83
(T.d.l) Woman without her man is a savage.
(T.d.2) Woman without her, man is a savage.
(T.d.3) Woman without her man, is a savage.
(T.d.4) Woman, without her, man is a savage.
(T.d.5) Woman without her man is a savage?
(T.d.6) Woman without her, man is a savage?
(T.d.7) Woman without her man, is a savage?
(T.d.8) Woman, without her, man is a savage?
(T.d.9) Woman without her man is a savage!
(T.d.10) Woman without her, man is a savage!
(T.d.l 1) Woman without her man, is a savage!
(T.d.12) Woman, without her, man is a savage!
The good examples of usage of ancient, intermediate and modern language
(T.h) are as follows:
(T.h.l) Book of "Sejarah Melayu" (The Malay Annals) ("Sejarah Melayu,"
1954) in the ancient Malay language
(T.h.2) Book of "Sejarah Melayu" (Shellabear, 1975) in the intermediate
Malay language
A few good examples are given below to show the effect of polysemy (T.j):
(T.j.l)Good
(T.j.2) Well
(T.j.3) Shell
(T.j.4) Lay
(T.j.5) Heart
84
6.2.4 Multimedia Type of Image
For the multimedia type of image, graphics, or picture, there are hard copy
and digital copy after the advent of computer technologies. The various images can
be categorized into eight groups (L1-L8) as follows:
(LI) Hard copy with simple black-and-white colours
(L2) Hard copy with complex black-and-white colours
(L3) Hard copy with simple multiple colours
(1.4) Hard copy with complex multiple colours
(1.5) Digital copy with simple black-and-white colours
(L6) Digital copy with complex black-and-white colours
(1.7) Digital copy with simple multiple colours
(L8) Digital copy with complex multiple colours
The first error level checking is very useful for digital copy of image.
Nevertheless, the tool of computing devices is required as an intermediate agent
between the human and the image. In addition to the first error level, the advanced
error of image as the second error level is important to be taken into consideration for
information mining. These semantic errors of image (La-Lh) may possibly occur as
listed below:
(La) Light source(s)
(Lb) Other light source(s)
(Lc) Optical properties of the viewed object
(Ld) Viewer
(I.e) Distance between the viewed object and viewer
(Lf) Medium conditions for the space of light of sight
(Lg) Relative position and velocity of light sources, viewed object and viewer
85
(Lh) Material decaying factor for the surface of the viewed object
The example of light source(s) (La) and optical properties of the viewed
object (Lc) are as follows:
(La.l) The star of Sun emits the visible light beam of white colour. It is to
note here that the basic colours of red, green and blue will compose
the white colour. The black colour will be the complement to the
white colour.
(Lc.l) The satellite of Moon is a viewed object reflecting the visible light
beam of white colour from the star of Sun. The human bare eyes will
see the Moon to be in yellow colour being the mixture of red colour
and green colour.
Hence, the multimedia type of image is not self independent for bit stream
and text, but it depends on the environment factors and viewer as well. In other
words, the information mining of image is not an objective issue but a subjective
issue.
6.2.5 Multimedia Type of Audio
For the multimedia type of audio or sound, it is either directly communicated
from the originator or indirectly communicated via a storage device. The audio data
storage can be either in the analogue form or the digital form. In the categorization
process, there are four groups (A.1-A.4) as follows:
(A.l) Aural sound (audible to human ears)
(A.2) Vocal sound (speech from human mouth)
(A.3) Environmental sound
(A.4) Music (Jacobs, 1998)
86
The first error level checking is basically over the bit stream of audio data
that is stored digitally after the advent of computer technologies. For the information
mining of audio data, it is important to consider the occurrence of second error level,
which has the semantic errors of audio (A.a-A.j) as listed below:
(A. a) Homograph where a spelling has many meaning
(A.b) Variation of a human voice like biological factor
(A.c) Pronunciation variation due to contextual factor
(A.d) Musical intonation
(A.e) Environmental factors and noises
(A.f) Situation of speaker
(A.g) Situation of listener
(A.h) Distance between the speaker and listener
(A.i) Medium conditions for the transmission of sound
(A.j) Relative position and velocity of speaker, listener and others
A few good examples are given below to show the effect of homograph (A. a):
(A.a.l)"Sea，， and "see"
(A.a.2) "Tree" and "three"
(A.a.3) "Plain" and "plane"
(A.a.4) "Flower" and "flour"
(A.a.5) "Five" and "fine"
For different musical instrument, the musical intonation will be different. The
examples of musical intonation (A.d) are as follows:
(A.d.l) Piano
(A.d.2) Guitar
87
(A.d.3) Saxophone
(A.d.4) Drum
Hence, the multimedia type of audio has lower accuracy rate than the
multimedia types of text and image in term of information mining. This is especially
true when it comes to the advanced error level of semantic error in the human-
machine, human- human, and human- machine-human communication systems.
6.2.6 Multimedia Type of Video
For the multimedia type of video or animation, it is in fact a combination of
text, image, and/or audio with the inclusion of time synchronization factor. Video is
in fact a word meaning "I see". The basic video is the body language as like the hand
signal, facial expression, gait, etc. This type of multimedia data becomes popular
after the advent of television technologies and computer technologies. There are
three main groups of video (6.1-6.3) as follows:
(V.l) Body language
(V.2) Analogue video data
(V.3) Digital video data
For the consideration of first error level of video, it is alike the error checking
process of the bit stream. For the second error level of video data, it includes all the
advanced error of semantic error as for multimedia types of text [(T.a) - (T.j)]， image
[(La) - (Lh)]， and audio [(La) - (Lj)]. Other than these presented semantic error, there
are other semantic errors of video data (V.a-V.c) as follows during the process of
information mining:
(V.a) Ambiguous meanings of the body language
(V.b) Time synchronization factor for text, image and audio
(V.c) Material decaying factor of the video storage media
88
The example of ambiguous meanings of the body language (V.a) is like the
polite protocol in welcoming the guest.
The example of time synchronization (V.b) is like the situation of mo vie- like
watching, where the sound system is not designed in good manner and the physical
distance creates multimedia differential error of image and audio.
In summary, the semantic error occurrences of video data have about 28
entries. Hence, the accuracy level for the multimedia type of video is lower than the
other types of multimedia in the process of information mining.
6.2.7 Duplication of Multimedia Data
After considering the semantic error occurrences for the various types of
multimedia data, it is hereby to study the ease of duplication for the multimedia. In
simple sentences, there are two types of multimedia data duplication as follows:
(D.l) Hard copy duplication directly usable by human
(D.2) Digital copy or others which require intermediate tool(s) for multimedia
data retrieval to be usable by human
Those in the subset of (D.l) are multimedia groups of text [(T.l)] and image
[(LI) - (L3)]. Meanwhile, those in the subset of (D.2) are multimedia groups of bit
stream [(B.l) - (B.3)]， text [(T.2)]， image [(1.4) - (1.8)], audio [(A.l) - (A.4)] and
video [(V.l) - (V.3)]. Hence, it can be concluded here that subset of (D.l) is more
human- friendly than the subset of (D.2).
6.2.8 Conclusion
In a nutshell, the multimedia error occurrences have been discussed for the
multimedia types or bit stream, text, image, audio, and video. The first error level
checking is not enough if we wish to increase the accuracy level of information
mining.
89
Subsequently, the advanced error at the second error level, which is semantic
error, is presented here for the various types of multimedia data. Also, the ease of
multimedia data duplication has been discussed.
It is hoped that Section 6.2 can help increase the accuracy level of
information mining during the multimedia communications. Expected fields for
implementation are science, commerce, history, archaeology, etc. One of the few
examples for possible practical implementation will be the decryption project for the
discovered manuscripts from the Dunhuang caves and other Silk Road sites
(International Dunhuang Project (IDP), No date).
6.3 Decrypting English Text Using Enhanced Frequency Analysis
Frequency analysis is the fundamental cryptanalytic technique besides brutal
force, threat, blackmail, torture, bribery, etc. (Schneier, 1996， 2000). Conventionally,
it is applied on monogram, bigram, and trigram together with anagramming
technique (Bauer, 2002). The efficiency of anagramming depends on the
cryptanalyst's depth of knowledge in different language features. Hence
anagramming is a trial-and-error approach due to huge deviation in different fields,
e.g. military, diplomatic, commercial, legal, and daily languages.
The frequency fluctuations among the normal frequency order of the
characters trigger the crossover problem for monogram frequency analysis.
Meanwhile, bigram and trigram frequency analyses are found to give little help.
Higher orders are unlikely to be useful. In year 1994， George Hart (1994) introduced
a novel frequency analysis approach based on word frequency to decode the
enhanced frequency analysis for systematic decryption of short English ciphertext.
The proposed protocol (Lee, Teh & Tan, 2006) in this Section 6.3 uses
combined techniques of monogram frequency analysis, keyword rules, and
dictionary checking. It is successfully tested faster on monoalphabetic substitution
cipher. Moreover, it solves the weakness of Hart's approach for complete failure
when neither one of the top 135 words is found in the ciphertext. Further
90
extrapolation is expected to be applicable to other cryptanalytic algorithms and
languages. Specifically here, the possible cracking of short password ciphertext can
be reflected.
6.3.1 Introduction
In the scientific world of secret writing, there are both the branches of
steganography using the hiding techniques and cryptography using the scrambling
techniques (Singh, 1999). To complement the cryptography, which is to reveal the
scrambled messages, the human culture achieves the first historical breakthrough in
the East via a brilliant mixture of linguistics, statistics, and religious devotion (Kahn,
1996a, 1996b). This new knowledge arena is called cryptanalysis by a brilliant
crypto logist William Friedman in 1920 (Wrixon, 1998). The etymology of this word
is from Latin "crypta", Greek "krypte" for "crypto" meaning "secret, hidden"; the
Greek "ana" for "ana" meaning "up, throughout"; and the Greek "lysys" for "lysis"
meaning "a loosing". In other words, cryptanalysis is the interception and solution of
a message by a third party. Both cryptography and cryptanalysis are collectively
known as cryptology.
6.3.2 Justification on Frequency Analysis
Among so many cryptanalytic techniques, frequency analysis or frequency
count is the most basic one other than brutal force, threat, blackmail, torture, and
bribery. The frequency analysis is in fact the anatomy of a language. According to a
book "Trattati in cifra" published in year 1470 and written by Leone Battista Alberti,
who is known as "Father of Western Cryptology", the aspect of cryptanalysis using
frequency analysis can be traced back to al-KindT, who is "The Philosopher of the
Arabs" and author of 290 books on medicine, astronomy, mathematics, linguistics
and music (Kahn, 1996a, 1996b; Bauer, 2002).
In 1987， the Arabic scientist al-KindT's treatise was discovered in the
Sulaimaniyyah Ottoman Archive in Istanbul and entitled "A Manuscript on
Deciphering Cryptographic Message" (Singh, 1999). It is believed that this
91
manuscript is the first ever known oldest description of cryptanalysis by frequency
analysis. The al-KindT's manuscript was written during the ninth century AD
containing discussions on statistics, Arabic phonetics, and Arabic syntax. Al-KindT's
explanation on the revolutionary system of cryptanalytic technique, i.e. frequency
analysis, is translated from Arabic language into English language (Singh, 1999) as
shown below. It best describes the fundamental operating principles of conventional
frequency analysis technique.
"One way to solve an encrypted message, if we know its language, is to
find a different plaintext of the same language long enough to fill one sheet or
so, and then we count the occurrences of each letter. We call the most frequently
occurring letter the 'first', the next most occurring letter the 'second', the
following most occurring letter the 'third', and so on, until we account for all
the different letters in the plaintext sample.
Then we look at the ciphertext we want to solve and we also classify its
symbols. We find the most occurring symbol and change it to the form of the
'first' letter of the plaintext sample, the next most common symbol is changed to
the form of the 'second' letter, and the following most common symbol is
changed to the form of the 'third' letter, and so on, until we account for all the
cryptogram we want to solve. "
Figure 6.1 Al-KindT's explanation on the revolutionary system of cryptanalytic
technique
All the other cultures then only knew about cryptography and not
cryptanalysis. Therefore, it is said that cryptology is born among the Arabs. In
addition to al-KindT's manuscripts, the Arabic knowledge of cryptology was fully set
forth in the section of a 14- volume encyclopaedia "Subh al-a ' sha" written by al-
Qalqashandi in 1412 (Kahn, 1996b). Al-Qalqashandi attributed most of his
cryptologic information to be from Ibn ad-Duraihim (1312-1361). Ibn ad-Duraihim-
Qalqashandi described the more analytical frequency analysis technique by using the
92
monogram first, then the two-letter word, three-letter word, and so on. Then the
concept of two-letter word gave rise to the concept of contacts of letters.
Thereafter, the Arabic cryptologic knowledge in finance, diplomacy, and
military fell for 250 years later. Nevertheless, the frequency analysis techniques
using monogram and contacts of letters developed by Arabians have become the
most universal, most basic cryptanalytic procedures. It is a prerequisite for
understanding all subsequent substitution cryptanalysis techniques (Schneier, 1996).
Almost at the same time of Ibn ad-Duraihim, the Western civilization began
the record of crypto logy around 1326 (Kahn, 1996b). Since then, the Western
cryptographic systems merged both of the code and cipher using substitution and
transposition for 650 years in both Europe and America. We call this as secret key
cryptosystem or classical cryptography (Kaufman, Perlman & Speciner, 1995).
During this period, the substitution cipher evolved from monoalphabetic substitution
cipher, to homophonic substitution cipher in 1401， to polygram substitution cipher
and polyalphabetic substitution cipher in year 1568 (Schneier, 1996). With the
advent of computer, the classical cryptography can easily be decrypted using the
frequency analysis and anagramming.
Subsequently, a new cryptographic field called as public key cryptosystem
( Stalling s, 2006b) or modern cryptography was introduced by Diffie and Hellman in
1976 (Diffie & Hellman, 1976). In practical life nowadays, both classical
cryptography and modern cryptography are used together under a hyond scheme, e.g.
IBM mainframes (Le， Matyas, Johnson & Wilkins, 1993)， to gain an optimization
between the key security and computational load. Public key cryptosystem is used
for exchanging the master keys. Meanwhile, the master keys are used for distributing
the secret session keys. The session keys will then be used under the secret key
cryptosystem. Now, most of the good cryptographic algorithms combine both of the
crypto systems with the mixed usages of substitution and transposition ciphers.
This is applicable and conveniently implemented with the handy help of
nowadays computing facilities. The only difference is just that instead of the 26
elements of alphabetical characters, now we have only two elements, i.e. binary '1'
and '0' (Schneier, 1996). Although it looks more complicated, it is basically the
93
same. It is just like the analogy of numerical computations between binary, decimal,
and hexadecimal numerical systems. Hence the frequency analysis remains as the
prevailing basic requisite for cryptanalysis techniques. This paper presents an
improved approach of modified frequency analysis upon the Hart's approach (1994)
for the decryption of English ciphertext. The improved cryptanalytical approach is a
ciphertext-only attack using the combined techniques of monogram frequencies,
keyword rules, and dictionary checking.
6.3.3 Drawbacks of Conventional Frequency Analysis
At a first look into the sentences that describe the frequency analysis on
monogram, the method shall work theoretically for sufficient long texts (Bauer,
2002). However, it is found in practical tests that the frequency distribution of
English language is a science fiction.
This is due to the fact that the English texts in different circumstances deviate
between fields, e.g. military, diplomatic, commercial, legal, literary, etc. Besides,
different person may use different words at different frequencies. In addition, an
energetic language is alive and always grows with time. The vocabulary is getting
more and more abundant. It may even undergo mutation due to the enrichment of
foreign words, e.g. paddy, spaghetti, kimono, etc.
Table 6.1 Frequency analyses of various sources [ascribe to Bauer (2002)]
Frequency Counts Descending Order
Sources
eaoid hnrst uycfg lmwbk pqxz
E.A. Poe 1843
etaoi nshrd lucmf wypvb gkqjx z
0. Mergenthaler 1884
etoan irshd lcfum pywgb vkxjq z
P. Valerio 1893
etaon isrhl dcupf mwybg vkqxj z
H.F. Gaines & O.P. Meaker 1939
etoan irshd lcwum fygpb vkxqj z
L.D. Smith 1943
etoan irshd lufcm pywgb vkxzj q
L. Sacco 1951
etaon irshd lucmp fywgb vjkqx z
D. Kahn 1967
etaon rishd lfcmu gpywb vkxjq z
A.G. Konheim 1981
etaoi nsrhl dcumf pgwyb vkxjq z
C.H. Meyer & S.M. Matyas 1982
94
Hence, these factors may cause the frequency fluctuations for the alphabetic
letters in the texts. The fluctuations may then create the crossover problems when we
try to match the empirical frequency counts with the ideal frequency counts. Table
6.1 above lists the old frequency analyses for English text with only 10,000 or fewer
letters adopted from Bauer (2002). It can be observed in Table 6.1 that the frequency
counts depend on the source file at certain deviation degree. The crossover problems
in frequency analysis can be ameliorated if the ciphertext is big. A huge ciphertext
can reduce the fluctuations of letters as practically tested by Bauer (2002) on German
texts. However, it is hard to have a chance to encounter with huge ciphertext.
6.3.4 Combined Methods of Frequency Analysis
To deal with these problems, Bauer (2002) proposed the cliques and partition
matching method. Cliques are groups of letters with almost equi-frequency that are
hard to be separated. This method is proven to be good on the conditions that the
cliques are not overlapping and having a clear gap between the members of a clique.
Hence, it normally works for intermediate- size ciphertexts and not the short
ciphertexts.
In addition, Bauer (2002) suggested other advanced frequency analysis
methods. The frequency of multigrams, such like bigram frequencies of 676 symbols
and trigram frequencies of 17,576 symbols, are found to be helpful in solving the
short ciphertexts that the cliques are not separated well by monogram frequencies.
However, higher orders of /i-gram frequencies with n>4 are almost helpless.
Other combined methods proposed by Bauer (2002) were word frequencies
between two spaces, position frequencies of a letter in a word, average word length,
and word formation by alternating and/or accumulating vowels and consonants.
Although there are numerous methods here, the only aim is to mechanize the
decryption of monoalphabetic simple substitution ciphers. Instead of huge and
intermediate texts, our aim also involves the short texts. Bauer's methods fade for
short texts.
95
For short cryptograms or ciphertexts, Hart (1994) proposed an efficient
algorithm using the language model. His method of [26!] partial permutation
searched through a manageable tree of word assignments. These words are from a
listing of 135 top-ranked words of modern American English in descending order
(Kucera & Francis, 1967). Firstly, starting from the smallest /i-letter words with same
length and same pattern of repeated letters if any repeat, we compare the n-letter
words of ciphertext with the 135-word dictionary to check for any possible
deciphering. The minimum value of n is 1 and the maximum value is 7.
"QUESTION" is ranked 358th. In other words, it means there is no 8-letter word in
the 135-word dictionary.
Nevertheless, there are some drawbacks in this method as well. While dozens
of words match the general 2-letter and 3-letter patterns, higher orders of n-letter
words after that are of low possibilities to have repeated letters. "THAT" is the only
case having the first and fourth letters repeated for 4-letter words. The small size of
135- word dictionary also causes the problem of missing letters like letters T， 'Q，，
'X，， and 'Z，. Hence, one may need to do anagramming via the trial- and-error method
to get the ciphertexts in proper meaning by inspection. Therefore, the approach to
increase the size of the word dictionary is helpful. It will help decipher the missing
letters from the word dictionary.
However, this method has no usage of grammatical information and relies on
word counts. In the special case of unusual letter distributions, this cryptanalytic
technique may be confused and ran. For the worst case if no plaintext is in the word
dictionary, it will definitely fail completely at a possibility of 2~ n for n-word texts
using a 135-word dictionary. Consequently, it is our purpose here to propose an
enhanced frequency analysis technique to mechanize the deciphering of this type of
ciphertext.
6.3.5 Proposed Method: Enhanced Frequency Analysis
In this proposed enhanced frequency analysis technique, it is a three
hierarchical approach. The monogram frequencies, keyword rules, and dictionary
96
checking are implemented one by one to mechanize the full deciphering of
monoalphabetic substitution cipher.
In the first step, the monogram frequencies are computed for all of the letters
in the ciphertext. Basically, at least the letters of 'E' and T' can be identified or
deciphered. During the second step, the keyword rules are conducted to decipher the
other unknown secret letters one by one. Each keyword rule is supposed to perform
only on cipher words that have only one secret letter. This secret letter will then be
deciphered. The sequence of the keyword rules is illustrated in Table 6.2 below. It is
to note here that some of the keyword rules may have the sequence swapped. Besides,
for letters that are still not yet deciphered in the second step, they will be identified in
the third step.
Table 6.2 Sequence of keyword rules for modified frequency analysis
Sq.
Letter
Keyword
Sq.
Letter
Keyword
1
to
11
u
our, out
2
n
no, on, not
12
c
can, could
3
f
of
13
w
we, was, with
4
or, for
14
g
go, get
5
a
a, at, are
15
1
all, will, could
6
d
do, and
16
b
be, but
7
h
he, the
17
y
by, may, yes
8
i
it, in, if
18
V
have, very
9
s
so, is
19
p
up, people
10
m
from, am, me
20
k
like, back, make
In the last step, there are at least four letters "JQXZ" that have to be
recognized by using the dictionary checking or inspection. If the ciphertext still has
lots of secret letters, dictionary checking will be in favour by having the cipher words
with the least number of secret letters tested first. If the ciphertext has very few
secret letters left as for alphabets "xjqz", then the inspection method is already
sufficient to decrypt them.
This three-tier approach has been tested on two short cryptograms with the
dimensions of 9006 letters for an excerpt from the storybook of "Arabian Nights: Ali
97
Baba and the Forty Thieves" (Penguin Popular Classics, 1997) and 2802 letters for
an excerpt randomly chosen from a daily newspaper article entitled "Bangsar". It is
found that both achieve very successful deciphering results at promising
computational time. The results are plotted in Figure 6.2. Hence, the deciphering
model to mechanize the cryptanalysis of monoalphabetic substitution cipher has been
successfully developed and tested.
It is found that this enhanced frequency analysis approach performs at more
systematic and faster decryption than the Hart's approach. Moreover, due to the
special combined properties of keyword rules and dictionary checking, the Hart's
approach weakness (1994) in facing total failure when neither one of the top 135
words are in the English ciphertext could be hindered.
e t o n f radhi smucwg lbyvpkj q x z
Sequence of alphabetic letters
Figure 6.2 Implementation of enhanced frequency analysis
6.3.6 Conclusion
In a nutshell, the cryptanalytic technique of enhanced frequency analysis has
been successfully developed by using the combined techniques of monogram
frequencies, keyword rules, and dictionary checking. The proposed three-tier
40 --
60 --
V
20 --
Excerpt from "Ali Baba & the 40 Thieves"
Excerpt from "Bangsar"
(%) uo^aAJ03p> IO W^IISOJSJ
98
approach in Section 6.3 mechanizes the cryptanalysis of monoalphabetic simple
substitution cipher. It is also discovered in the research that this enhanced approach
shows improved performance from the Hart's approach in the context of English
texts. The improvements are faster ciphertext decryption and the avoidance from the
chances of total railure when none a single top 135 words is within the ciphertext.
In future, the keyword rules may be refined to have a faster deciphering speed
by rearranging the sequence order and/or the selected keyword for specific letters.
Furthermore, finer and more advanced algorithms could be established from here to
decipher the other advanced substitution and transposition ciphers, e.g. homophonic
substitution cipher, polygram substitution cipher, polyalphabetic substitution cipher,
and simple columnar transposition cipher.
It is our wish that other advanced ciphers like Playfair Cipher could be
decoded. Moreover, there are possibilities that this model could be applied to other
languages as well. The advent of Internet has created an e-life society that is not only
multidisciplinary but also multilingual. Perhaps, the Scrabble Players Dictionary
('The Official Scrabble," 2002) may become another source of idea towards the
direction of further improvement. Specifically here, Section 6.3 reflects some facts
on the possible cracking of short password ciphertext.
6.4 Passphrase with Semantic Noises and a Proof on Its Higher Information
Rate
Key size becomes very important to a cryptographic algorithm according to
Kerckhoff s law where a civilian crypto system shall depend fully on key secrecy.
Currently, there are 4 passphrase generation methods: Sentence, acronym, diceware,
and coinware. Unicity distance is the minimum ciphertext size for unique ciphertext
decipherability when number of spurious keys is zero. A key with size less than
unicity distance is good where there are spurious keys which allow a protection
method using limited unsuccessful logins. Here (Lee & Ewe, 2007d)， stronger forms
of passphrases using textual semantic noises like punctuation marks, mnemonic
substitution, misspelling, and associative morphing, whicn improve the key entropy,
are proposed. An ASCII mutual substitution table is presented together with its proof
99
on information rate increment. Higher information rate has lower redundancy, and
hence bigger unicity distance ensures encrypted keys the short cryptogram in a key
vault, like Password Safe, cannot be cracked within certain limited login attempts.
6.4.1 Introduction
Key is one of the four main groups of entity authentication for identification.
These four groups are "something known", "something possessed", "something
inherent", and "someone known" (Menezes, Oorschot & Vanstone, 1996). Key is a
secret only known to the authenticated entity, where its low cost, mobility, and wide
compatibility, makes it to be the most popular authentication method.
Kerckhoff s law is applied in civilian crypto system, where the strength of a
crypto system is fully dependent on the key secrecy (Schneier, 1996). In other words,
key size is the main factor of a cryptographic algorithm. Short key is called password
and long key is called passphrase. FTPS PUB 112 (Federal Information Processing
Standards Publication 112) dated 30 May 1985 on password usage (NIST, 1985b)
defined password to be a key with a length of 4 to 8 characters and passphrase as a
key with a length of 9 to 64 characters.
To estimate the key entropy, NIST (National Institute of Standard and
Technology), USA, published an electronic authentication guideline (Burr, Dodson
& Polk, 2004， 2006; Burr, Dodson, Perlner, Polk, Gupta & Nabbus, 2008). However,
the estimation of user-chosen key entropy is based on Shannon's information theory
(Shannon, 1948) and English information rate (Shannon, 1951; Cover & King, 1978)
using 2o English alphabets plus one space character. For more accurate figures, the
English information rate is in fact shall be based on 95 ASCII printable characters,
where its limiting (infinite- length history) conditional entropy is available in (Brown,
V. J. D. Pietra, Mercer, S. A. D. Pietra & Lai, 1982; Tsou, Lai & Chow, 2004).
A good key shall be strong and memorizable. A strong key has high entropy
and high randomness. Meanwhile, a memorizable key has reasonable secrets to be
remembered. Weak key is not in favourite (Klein, 1990; Spafford, 1992a, 1992b).
Random key is strong but not memorizaole. Hence, this thesis proposes keys with
100
balanced features of strength and memorability (Kurzban, 1985; Bugaj, 1996; Yan,
Blackwell, Anderson & Grant, 2004).
Due to the long key size demand for symmetric key algorithm and
asymmetric key algorithm (E. Barker, W. Barker, Burr, Polk, & Smid, 2007a, 2007b;
Gehrmann & Naslund, 2005， 2006， 2007)， password is no longer enough and
passphrase (Schneier, 1996; Menezes, Oorschot & Vanstone, 1996; Stalling s, 1995)
is needed. For the popular email encryption software, PGP (Pretty Good Privacy)
version 9.0， the allowed key size comes to a maximum of 255 characters (PGP
Corporation, 2006). Nowadays, many modern operating systems support a maximum
key field of 255 characters.
For the entry of passphrase, there are currently four input methods: Sentence,
acronym, diceware (PGP Corporation, 2006) and coinware (Lee & Ewe, 200oj. In
this section, a stronger form of passphrase is proposed to have more spurious keys by
using the semantic noises or semantic errors (Lee, 2005a) like punctuation marks,
misspelling, mnemonic substitution, and associative morphing (Bugaj, 1996).
6.4.2 Key Sizes and Passphrases
The minimum symmetric and asymmetric key sizes for different protection
periods are given by (E. Barker, W. Barker, Burr, Polk, & Smid, 2007a, 2007b;
Gehrmann & Naslund, 2005， 2006， 2007) as in Tables 3.1 and 3.2.
The ASCII entropy is 6.57 bits/letter. Hence, many key sizes are challenging
the human memorizability limit. For asymmetric key algorithm before the MePK ᄂ，
the long key sizes make it impossiole to be memorizable in ASCII encoding code
and require encrypted private key stored in the computing device. A user has to
remember the shorter symmetric key used to protect the encrypted private key.
Hence, asymmetric key algorithm before the MePKC is normally having the
portability and mobility proolems of private key. From Table 3.1， the minimum key
size for smallest general purpose level with a maximum of 2-year protection or
before year 2010 is 80 bits. If ASCII is used, it needs 13 characters to fulfil the key
size requirement. Hence, passphrase shall be used as compared to password.
101
There are four types of passphrase generation methods (Lee & Ewe, 2006):
Sentence, acronym, diceware (PGP Corporation, 2006)， and coinware (Lee & Ewe,
2006). Sentence-type and acronym-type passphrases are subject to computational
analysis of word frequency distribution. Meanwhile, diceware and coinware are
immune to the computational analysis due to the feature of random word selection
from dictionary and self- created word list(s).
Sentence-type passphrase uses an entire phrase or full sentence to form a key.
Acronym-type passphrase applies abbreviation of first, second, last, etc., letters of
each word in a sentence. Diceware uses dice to choose a word from an ordered word
list. The word list can be in any language and based on senary or base- 6 numeral
system. Coinware is similar to diceware except it uses coin to select from a word list
that can be monolingual, bilingual, or multilingual. It is especially efficient for word
list in binary, octal, and hexadecimal orders. There are readily built word lists for
Han characters in Unicode-encoded CJK languages.
6.4.3 Unicity Distance and Passphrases with Semantic Noises
Capitalization and permutation are the prior arts to increase the passphrase
entropy (Stallings, 1995). Mnemonic substitution and associative morphing are other
forms of prior arts. The latter two methods are presented in very brief manner
without a proof (Bugaj, 1996). Here, these four methods are generalized together
with another two methods from the author's research project, i.e. punctuation marks
and misspelling, as passphrase with semantic noises, and propose a user template of
ASCII mutual substitution table, which comes together with a proof on the
information rate increment.
Passphrase with semantic noises has higher information rate (r)， lower
redundancy (D) (Ritter, 2001; Wikipedia Contributors, 2008f), and hence bigger
unicity distance (n ). Unicity distance (Stinson, 2002; Wag staff, 2003) is the
minimum ciphertext size for unique decipherability of ciphertext given sufficient
decryption time, when number of spurious keys (F) is zero. The larger the difference
between the unicity distance and key length, the more the spurious keys, and the
stronger is the protection method using limited login attempts. Special key
102
management algorithms allow multiple site keys (aka slave keys) to be created from
a master key (Yee & Sitaker, 2006; Lee & Ewe, 2007a). This further permits each
short cryptogram of keys in a key vault like Password Safe to be encrypted by
different slave keys. The decoding of short cryptogram has been studied by Hart
(1994). Faster decryption can be achieved using an enhanced frequency analysis (Lee.
Teh & Tan, 2006). The relationships of information rate, absolute rate like random
signal (R), key entropy (H(K))， ciphertext size (n)， and umcity distance are given in
Equations (6.1-6.4).
D = R_r (6.1)
F>2 H(K) " nD -l (6.2)
n > H(K) / D when F = (6.3)
r| ^> D| ^> Ft ,n t. (6.4)
For instance, English text has r = 1.3 bits/letter for 27 symbols (a, b, c, ...， z，
space) (Shannon, 1951; Cover & King, 1978). The redundancy is R = 4.75 bits. If
AES-128 is used, H(K) = 128 bits. Hence, n = 128/3.45 = 38 characters. For 95
ASCII printable characters, the upper bound of r based on English language becomes
1.75 bits/letter (Brown, V. J. D. Pietra, Mercer, S. A. D. Pietra & Lai, 1982). The
revised n = 128/(8-1.75) = 21.
Below are the examples of passphrases with semantic noises using
punctuation marks, misspelling, mnemonic substitution, associative morphing,
capitalization, and permutation. Punctuation mark is the easiest. A user is encouraged
to embed semantic noises for all types of passphrases: Sentence, acronym, diceware,
and coinware. Wing ate and Sinden (2002) have written a book on how to create short
text messages. SMS (Short Message Service) language (Wikipedia Contributors,
2008ag) and text messaging (Wikipedia Contributors, 2008ah) are other good
references. The presence of these spurious keys is very useful for the case of limited
login attempts, where unique cryptanalysis is impossiole.
103
Actual key: Woman without her man is a savage.
Semantic noises: Punctuation marks and permutation {
Woman without her, man
is
a savage.
Woman without her man,
is
a savage.
Woman, without her man
is
a savage.
Woman without her, man
is
a savage?
Woman without reh man,
si
a savage?
Woman, without reh man
si
a savage?
Woman without her man
is
a savage!
Woman without her, man
is
a savage!
Woman without reh man,
si
a savage!
Woman, without reh man
si
a savage!
}
Actual key: To be, or not to be: That is the question.
Semantic noises: Misspelling and capitalization {
To be? or not to be? That is the question?
To be, or not to be? that is the question!
TO 6e? Or nOt tO 6e? th@t i5 the que5ti0n?
To be! Or not to be! That is the question!
TO BE! OR NOT TO BE! THAT IS THE Question!
To we, or not to we: That is the question.
To me, or not to me: That is the question.
To be, of not to be: That is the question.
}
104
Actual key: Ballon, Address? Atmel. -Star-
Semantic noise: Mnemonic substitution {
B@! ！ On, Address? Atmel. -Star-
Ballon, Address? @mail. -Star-
}
Semantic noise: Associative morphing {
Ballon, +++re$$? Atmel. -Star-
Ballon, Address? Atmel.
B@!!0n， +++re$$? @mail. ~****~
}
Table 6.3 ASCII mutual substitution table
aA
bB
cC
dD
eE
fF
gG
hH
il
jJ
A
6
<
ol
3
1=
9
l-l
I
？
kK
1L
mM
nN
oO
pP
qQ
rR
sS
tT
l<
1
TVI
TV
lo
&
1-
5
+
uU
vV
wW
xX
yY
zZ
1
2
3
[J
\/
VV
><
7
2
1
Z
E
4
5
6
7
8
9
+
*
1
h
S
b
L
B
g
t
X
1
%
[
]
(
)
<
>
o/o
eq
[
]
<
>
(
)
I
#
$
&
i
n
m
Q
*
dot
？
@
\
A
1
space
j
at
A
\
1
A v
CamelCase
105
6.4.4 Proof of Higher Information Rate
Here, an ASCII mutual substitution table is presented as user template to
create passphrase with semantic noises. A user can modify any mutual substitution of
these ASCII characters in Table 6.3. CamelCase (Wikipedia Contributors, 2008g)
makes compound words or phrases in which the words are joined without spaces and
are capitalized within the compound. The ASCII substitution is a token with one or
more characters. The probability of the initial token letter is used as the token
probability, where the difference is assumed to be small and negligible.
The upper bound of information rate (r) (Shannon, 1951; Cover & King, 1978)
is given by Equations. (6.5-6.6)， where q N t is the probability for predictor to discover
the correct letter following a sequence of N-l symbols in i guesses, i indexes one of
the 95 ASCII printable characters.
qf =!>C /いん，…， ん— いん) (6.5)
Z^f -^ + i)log 2 i<r< -|>デ bg 2 qf (6.6)
Due to the mutual substitution of ASCII printable characters, 9^ q , becomes
about 48 pairs. Every two different q\ with different probabilities are paired to share
the same probability. Let one of these pairs has probabilities A and B before mutual
substitution, and probability C after mutual substitution. Let other ASCII characters
have a combined probability D, where (A + B + D = 1) and (A + B)/2 = C. The
inequality Equation (6.7) can be proven using differentiation of calculus dy2/dA in
Equation (6.11) as derived from Equations (6.8-6.10). yi is a constant and yi has an
absolute minimum value equal to yi at A = B = (1 - D)/2. The other two critical
points, A = and 1 - D, share the same absolute maximum value. Table 6.4 shows
how to determine the critical points of function y2.
_ 2C * log 2 C > _ A * log 2 A _ B * log 2 B (6.7)
C 2C < A A B B (6.8)
((1 - DV2) 1 " < A A (1 - D - A) 1_D " A (6.9)
106
y 1 = ((l-D)/2) 1D
y 2 = A A (l-D-A) 1DA
(6.10)
(6.11)
Inequality Equation (6.6) can be further extended to three or more mutually
substituted ASCII characters for higher information rate increment. The best case is
all the ASCII characters can be mutually substituted, which creates the highest
information rate like absolute rate the random signal, where unicity distance will
become infinite. However, this is just an ideal dream. What we can do is to approach
the dream as close as possible. As in Equation (6.4)， higher information rate has
more spurious key.
y 2 ，(A) = dy 2 /dA = y 2 In (A / (1 - D - A)) (6.12)
y 2 ，，(A)= d 2 y 2 /dA 2 (6.13)
y 2 ，，(A) = (dy 2 /dA) * In (A / (1 ᅳ으 A)) + y 2 * (1/A + 1/(1 ᅳ으 A)) (6.14)
When y 2 ，(A) = 0， then A = (1 - D)/2. (6.15)
Other critical points are A = and (1 -D). (6.16)
Since y 2 ，，(A = (1 - D)/2) = +ve，
then A = (1 — D)/2 is an absolute minimum value. (6.17)
y 2 (A = 0) = y 2 (A=l_D) (6.18)
Table 6.4 Determination of the critical points of function y2
Critical Points
A =
A = (l -D)/2
A= 1 _D
Leftofy 2 '(A)
-ve
+ve
Right ofy 2 '(A)
-ve
+ve
Types of Extreme
Values
Absolute
Maximum
Absolute Minimum
Absolute
Maximum
6.4.5 Conclusion
Here, stronger form of passphrase is proposed using semantic noises
generalizing the punctuation marks, capitalization, permutation, mnemonic
107
substitution, associative morphing, and misspelling. Passphrase with semantic noises
has higher information rate, bigger unicity distance, and more spurious keys, which
strengthens the login protection with limited attempts. In addition, an ASCII mutual
substitution table and its proof on information rate increment is provided.
6.5 Information Rates and Unicity Distances
Below are the denotation to show the information rates and unicity distances
of ASCII-based 2D key, Unicode-based 2D key, black-and-white multilingual key,
and colourful multilingual key. Unicity distance is the minimum size of ciphertext
for the unique decipherability of ciphertext when given enough time.
n = unicity distance (6.19)
F = number of spurious keys (6.20)
K = key (6.21)
H(K) = key entropy (6.22)
D = R-r (6.23)
where D = redundancy, R = absolute rate like random signal, r = information rate
F>2 H(K) — nD _l (6.24)
n > H(K) / D when F = (6.25)
For limiting conditional entropies of English language (Shannon, 1951;
Cover & King, 1978; Brown, V. J. D. Pietra, Mercer, S. A. D. Pietra & Lai, 1982;
Yannakoudakis & Angelidakis, 1988), Chinese language (Fossum, No date; Vines &
Zobel, 1998; Zhang, Xu & Huang, 2000; Tan & Yap， 2001; Tsou, Lai & Chow， 2004;
Li, Hu， Wang & Dai, 2005), Malay language (Tan, 1981)， and various languages
(Behr, Fossum, Mitzenmacher & Xiao, 2002， 2003)， these are the references.
English = 1.3 bits/letter (26 alphabets + space), 1.75 bits/ASCII character (6.26)
Chinese = 4.0462, 4.1， 4.5， 5.17 bits/character from four different studies (6.27)
108
For the information rate and unicity distance of ASCII-based 2D key:
D = 8- 1.75 = 6.25 (6.28)
n (128-bit key) = 128/6.25 = 20.48 (6.29)
n (256-bit key) = 256/6.25 = 40.96 (6.30)
Number of required ASCII characters (128-bit key) = 128/6.57 = 19.48 (6.31)
Number of required ASCII characters (256-bit key) = 256/6.57 = 38.96 (6.32)
For the information rate and unicity distance of Unicode-based 2D key:
D= 16-4.1 = 11.9 (6.33)
n (128-bit key) = 128/11.9 = 10.76 (6.34)
n (256-bit key) = 256/1 1.9 = 21.51 (6.35)
Number of required Unicode symbols (128-bit key) = 128/16 = 8.00 (6.36)
Number of required Unicode symbols (256-bit key) = 256/16 = 16.00 (6.37)
For the information rate and unicity distance of black-and-white multilingual key:
R= 17 + 3 (6.38)
r = 4.1 + 3 (6.39)
D = 20- 7.1 = 12.9 (6.40)
n (128-bit key) = 128/12.9 = 9.92 (6.41)
n (256-bit key) = 256/12.9 = 19.84 (6.42)
Number of required image selections (128-bit key) = 128/20 = 6.40 (6.43)
Number of required image selections (256-bit key) = 256/20 = 12.80 (6.44)
Number of required image selections (236-bit key with 20-bit key strengthening)
= 236/20= 11.80 (6.45)
109
For the information rate and unicity distance of (16+1) colourful multilingual key:
R= 17 + 3 + 8
(6.46)
ri = 4. 1 + 3 + 8 (colours encoded as digits)
(6.47)
or r 2 = 4.1 + 3 + 1.75 (colours encoded as alphabets)
(6.48)
II
00
II
o
ᅳ s
II
00
bo
ᄂ /i
II
1 ~ '
ᄂ n
(6.49)
n i (128-bit key) = 128/12.9 = 9.92
(6.50)
n i (256-bit key) = 256/12.9 = 19.84
(6.51)
n 02 (128-bit key) = 128/19.15 = 6.68
(6.52)
n 02 (256-bit key) = 256/19.15 = 13.37
(6.53)
Number of required image selections (128-bit key) =
128/28
= 4.57
(6.54)
Number of required image selections (256-bit key) =
256/28
= 9.14
(6.55)
Number of required image selections (236-bit key with 20-bit key strengthening)
= 236/28 = 8.43
(6.56)
If the numbers of required ASCII/Unicode characters and image selections (T)
are less than the respective unicity distance, then there will be no unique
decipherability. The more spurious keys, the better it is. For higher information rate
and smaller T， then there are more spurious keys when T < no. Setting an attempt
limit to open a key vault using a master key can then become a possible method of
protection. Equations (6.57-6.60) show the required formulas to estimate the unicity
distance and T.
Let T = numbers of required ASCII/Unicode characters
and image selections. (6.57)
F > 2 H(K) — TD — 1 when n = T， T く n (6.58)
T > (H(K) - log 2 (F + 1)) / D = n — (log 2 (F + 1))/D (6.59)
110
log 2 (F + 1) > D(n - T)
(6.60)
Moreover, to know the information content of selected images as secret key
in some of the big secret creation methods, please refer to Tavakoli (1991).
Ill
CHAPTER 7 MULTIHASH KEY
7.1 Introduction
There are lots of situations that require a user to have many online and offline
accounts. Examples of online and offline accounts are login access and file
encryption, respectively. For safer security, a secret cannot be re-used to avoid
password domino cracking effect (Ives, Walsh & Schneider, 2004), where an
attacker starts the password cracking process from the weakest link. However,
according to R. Kanaley (2001)， an Internet user manages an average 15 keys on a
daily basis. Yet in another survey by Adams and Sasse (1999)， a user can only be
expected to handle 4 to 5 unrelated and regularly used keys. For user's unique keys
without the constraint of relevancy, Florencio and Herley's survey (2007) reported
an average 6.5 keys, repeated 3.9 times each for 25 accounts and typing 8 keys daily.
Hence, there is a memory burden to the user unless these secrets are written down
somewhere. However, important password the secret is discouraged to be jotted
down somewhere.
7.2 Multiple hashes of single key with passcode for multiple accounts
A human' s e-life needs multiple offline and online accounts. It is a balance
between usability and security to set keys or passwords for these multiple accounts.
Password reuse has to be avoided due to the domino effect of malicious
administrators and crackers. However, human memorizability constrains the number
of keys. Single sign-on server, key hasmng, key strengthening, and petname system
are used in the prior arts to use only one key for multiple online accounts. The unique
slave keys (aka site keys) are derived from the common master secret and specific
domain name. These methods cannot be applied to offline accounts such as file
encryption. New method and system are invented to be applicable to offline and
online accounts. It does not depend on http server and domain name, but numeric 4-
digit passcode, key hashing, key strengthening, ana hash truncation. Domain name is
only needed to resist spoofing and phishing attacks of online accounts.
112
7.2.1 Introduction
For friendly environment, cost effectiveness, and efficiency, human
civilizations are heading towards a paperless and electronic society. Every human is
getting numerous offline and online accounts. These accounts require authentication
to gain system access. There are four types of authentication approaches: Secret,
token, biometrics, and introducer.
Secret is about something you know like password or key. Token is about
something you have like smart card. Biometrics is about something you are like fin-
gerprint. Introducer is whom you know. For the sake of cost and compatibility, secret
in the form of key is the most popular authentication approach.
According to Forrester Research (Kanaley, 2001)， an active Internet user
manages an average of 15 keys on a daily basis. Most people, who are majority-wise,
not using the password management tools, either maintain the same key for all the
accounts, write down different keys for different accounts, or keep closely related
keys for various accounts. These are all poor password management practices.
The HTTP basic authentication protocol (even over SSL) (Franks, Hallam-
Baker, Hostetler, Lawrence, Leach, Luotonen & Stewart, 1999) allows a server to
know the key of each account. Tms causes possible malicious server attacks from the
administrators and crackers. The server may be untrustworthy or compromised.
For another HTTP specification, i.e. HTTP digest authentication protocol,
challenge-response protocol is used (Franks, Hallam-Baker, Hostetler, Lawrence,
Leach, Luotonen & Stewart, 1999). The server can still see the clients' keys. Since
the response from a client to a server is not specific to the server, HTTP digest au-
thentication protocol is vulnerable not only to malicious server attacks, but password
file compromise attacks, spoofing attacks, and phishing attacks.
If a key is reused, the success of an attack on an account in a weak system
may cause a strong system to be compromised. This password reuse can trigger a
domino effect from the weakest system to the strongest system (Ives, Walsh &
Schneider, 2004).
113
Therefore, every key has to be uniquely set for each account, regardless of
weak or strong system, to get rid of the risk when one system is compromised.
However, according to (Adams & Sasse, 1999), users can only be expected to cope
with a maximum of four or five keys that are unrelated and regularly used. When key
relevancy is allowed, a user can cope with average 6.5 unique keys (Florencio &
Herley, 2007). This reflects the need to balance the usability and security.
To address this problem, some key management tools are invented. These
tools allow users to remember only one master secret as master key and assign
unique keys to multiple accounts. They allow users either to choose their own master
key and then store the site keys (aka slave keys) somewhere safe, or to assign fixed
keys to each website that can be computed whenever they are needed.
The examples of the first approach are Password Safe and Windows Live ID.
The examples of the second approach are LPWA (Lucent Personal Web Assistant),
HP Site Password, Password Multiplier, SPP (Single Password Protocol), PwdHash,
and Passpet. A special example using the hybrid approach is CPG (Compass
Password Generator).
Password Safe is a password vault that can be used for offline and online
accounts. However, its mobility is low due to the requirement to have a safe storage
for multiple keys encrypted by a common master key. Another form of solution for
online accounts only is to use a single sign-on server and its proxy servers. Microsoft
Windows Live ID (aka Microsoft Passport Network) is one of these examples. Its
weaknesses are single point of failure and high cost or integration.
Another method to reduce the memory burden of online account passwords
uses key hashing and key strengthening (aka key stretching) of a master key
concatenated with a domain name and optional username. Exemplary applications of
this method are (i) LPWA (Lucent Personal Web Assistant) (Gabber, Gibbons,
Matias & Mayer, 1997); (ii) HP Site Password (aka System-Specific Passwords or
Site-Specific Passwords) (Karp & Poe， 2002; Karp, 2003); (iii) Password Multiplier
(Halderman, Waters & Felten, 2005); (iv) PwdHash (Ross, Jackson, Miyake, Boneh
& Mitchell, 2005); and (v) Passpet (Yee & Sitaker, 2006).
114
There is also a method using unique random number assignment to different
online accounts called CPG (Compass Password Generator) (aka Common Password
Method) (Luo & Henry, 2003). Yet there is another method using the key hashing of
one-time ticket, server name, and master password to generate different site keys
(aka slave keys) called SPP (Single Password Protocol) (Gouda, Liu, Leung & Alam,
2005).
All these methods of single master key generating multiple site keys or slave
keys apply only to online accounts having a domain name. Its weakness is a change
of master key requires all the accounts to be updated one by one, which is required
by some key management strategies.
For offline account, the current prior art uses a password vault to store all the
unique passwords the secret. These password vaults can be simply an encrypted
spreadsheet or document file, or application software like Password Safe by Bruce
Schneier [URL: http://www.schneier.com/passsafe.html]. The disadvantage of
password vault is its low mobility and danger of disclosing the ciphertext of
password vault to the public domain. Hence, there exists a need to have a method to
generate multiple slave keys of online and offline accounts from a master key, and
yet an individual slave key can be changed without changing the master key and
other slave keys.
With the realization of big memorizable secret for cryptographic,
information- hiding, and non-cryptographic applications, especially MePKC, there
are even more types of offline accounts like asymmetric private key, stego-key,
symmetric watermarking key, asymmetric watermarking private key, and PRNG
seed. Among them, for MePKC cryptographic applications like encryption, signature,
authentication, key exchange, and other schemes, different schemes require a
different pair of asymmetric key pair, by the technical and law requirements to have
a safer electronic information society. Hence, there exists a need to generate multiple
private keys as slave keys from a common memorizable master key.
The present invention (Lee & Ewe, 2007a) can be applied to offline and
online accounts with good mobility. Domain name is not necessary but optionally
needed to resist phishing attacks and spoofing attacks. A single sign-on server is also
115
not needed. The required components are numeric 4-digit passcode, key hashing, key
strengthening, and hash truncation.
To allow diversity of site keys from a single master key, there are two
optional entries: Username ID and domain name (or website) URL. Domain name
that is also used to resist phishing attacks can be replaced by adopting an anti-
phishing tool. In other words, the proposed new method and system can be used
together with an anti-phishing tool.
These anti-phishing tools are SpoofStick, Netcraft Toolbar, Earthlink Toolbar.
SiteKey, DSS with SRP (Dynamic Security Skins with Secure Remote Password
Protocol), Petname Tool, TrustBar, and Passpet (Yee & Sitaker, 2006).
7.2.2 Related Works
Here, the prior arts of key management tools are discussed, where a single
key can be used for multiple accounts, in a deeper context. Anti-phishing tools will
not be discussed. Accounts are divided into two types: Offline and online. Offline
accounts have no domain name while online accounts have domain name. Example
of offline accounts is file encryption; whereas example of online accounts is email.
Password Safe is an application software originally developed by Bruce
Schneier [URL: http://www.schneier.com/passsafe.html]. It uses the Twofish
encryption algorithm to protect the stored passwords by a master password. Users
need only to remember one master password to access multiple passwords. Its
mobility depends on the available password database. It can be used for both offline
and online accounts, but cannot resist spoofing attacks and phishing attacks.
Windows Live ID (No date) is also known as "Microsoft Passport Network"
[URL: http://www.passport.net]. Users need a master password to sign on a central
server. This central server will authenticate users for multiple servers which have
joint network. Besides single point of failure, it has high cost or integration. Some
security loopholes are reported (Kormann & Rubin, 2000). It can be used for online
accounts only, but can resist phishing and spoofing attacks.
116
LPWA (Gabber, Gibbons, Matias & Mayer, 1997; Matias, Mayer &
Silberschatz, 1997) uses key hashing of master password and domain name to
generate a specific site password via a server. It has single point of failure but not the
high cost of integration. However, the malfunction of central authority will mean the
breakdown of all services. It can be used for online accounts only and can resist
phismng and spoofing at-tacks. Nowadays, it has stopped providing the services.
HP Site Password (Karp, 2003; Karp & Poe, 2004) is also called "System-
Specific Passwords" or "Site-Specific Passwords". A master password and a system
name are concatenated, hashed using MD5 (Rivest, 1992) and converted into Base64
encoding (Borenstein & Freed, 1992) to get a site password. It is not centralized
using a server but operates as stand-alone application in the terminal computers. It
can be used for online accounts only and cannot resist phishing and spoofing attacks.
It is important to note here there were few successful collision attacks over
the MD5 in the years 2004-2006 (Wikipedia, 2008r). The successor of MD5, which
is SHA-l, is also discovered to be subject to collision attacks on its reduced version
in the years 2004-2006 (Wikipedia, 2008w). Consequently, NIST announced that
SHA-l would be phased out by the year 2010 in favour of SHA-2 variants: SHA-224,
SHA-256, SHA-384, and SHA-512 (NIST, 1995a, 2002b; 2007b; Lilly, 2004).
CPG (Luo & Henry, 2003) is also called "Common Password Method". It
assigns unique random numbers to different website accounts. The random number is
hashed using MD5 and converted using a binary-to-text transform to generate a
specific password for multiple accounts. The random number is encrypted and stored
in an account server or proxy server. When a user needs to access a specific account,
the encrypted random number is retrieved from the server, decrypted, hashed, and
converted into a specific password to authenticate the access. Therefore, it has the
weakness of single point of railure, but does not involve the high integration cost like
LWPA. It is for online accounts only and can resist phishing and spoofing attacks.
Password Multiplier (Halderman, Waters & Felten, 2005) uses key hashing
and key strengthening. There are two levels of hash iterations using the inputs of
username, master password and site name. Both the numbers of hash iterations are
fixed for 100 seconds and 1/10 second, respectively. It is a stand-alone application
117
without using a server and implemented using browser extension to Mozilla Firefox.
It can be used for online accounts only and can resist phishing and spoofing attacks.
SPP (Gouda, Liu, Leung & Alam, 2005) is also a stand-alone application. It
applies the techniques of challenge-response protocol, one-time server- specific ticket
and key hashing using MD5 or SHA-1. The site password is hashed from the one-
time ticket, server name, and master password. The one-time ticket and site password
will be updated after every login access. It can be used for online accounts only and
can resist phishing and spoofing attacks.
PwdHash (Ross, Jackson, Miyake, Boneh & Mitchell, 2005) is implemented
using browser extensions to Mozilla Firefox, Internet Explorer, and Opera. Its key
hashing inputs the domain name of remote site into a pseudo-random function
controlled by user's master password. The domain name acts as a hash salt. It can be
used for online accounts only and resist pmshing and spoofing attacks.
Passpet (Yee & Sitaker, 2006) is also implemented using browser extension
to Mozilla Firefox. It applies the techniques of petname system, key hashing, key
strengthening, and UI customization. Petname system is a naming system possessing
the properties of globality, security and memorizability (Wikipedia, 2008k). It is
used for anti-pmshing attacks. Key hashing and key strengthening in Passpet are
alike the Password Multiplier using the SHA-256, except that its first level of hash
iterations is flexible in amount allowing updates according to the computer
technology advancement without changes of software. It uses local storage for login
access via a fixed machine, and remote storage in a server for login access with
mobility feature. The remote server stores the first level of hash iterations and site
label file that is encrypted from the site label list. Due to the dependency of server for
newly used machines, Passpet has some risks of single point of failure. However,
there is no high cost of integration. It can be used for online accounts only and can
resist phishing and spoofing attacks.
7.2.3 Basic Model of Multihash Key
The proposal here requires users to remember an at least 128-bit master key
and a numeric 4-digit passcode. Tms method and system is named "multihash key".
118
The passcode is used together with key hashing, key strengthening (Manber,
1996; Abadi, Lomas & Needham, 1997, 2000; Kelsey, Schneier, Hall & Wagner,
1997) and hash truncation to generate exemplary 20 unique hashes at 20 security
levels for 20 accounts. Each security level has one account. These hashes are site
keys. All the security levels are ranked from the highest security (#1) to the lowest
security (#20). This is because knowing the multihash key at the higher level can
reveal the multihash key at the lower level, but not the reverse.
From Kanaley's survey (2001)， 20 accounts are set since an active Internet
user manages an average of 15 keys daily. Five accounts are added by assuming that
there are five offline accounts. The number of accounts can be increased by changing
the settings or remembering another pair of (master key, passcode).
There are three pseudo-codes for multihash key to show how the method and
system work: Determination of hash iterations of multiple security levels, generation
of multihashes as site keys, and changes of key pair (master key, passcode).
As an example, Figure 7.1 shows the determination of 20 security levels via
the experiments to locate the lower bound &ᄂ and upper bound bu for 1 -second hash
iterations for an old computer that is slow but still popular. Each security level is
partitioned by 2 8 .
1400
V
Settings to determine the lower and upper bounds of 1 -second hash iteration:
(1) b L = lower bound for 1 -second hash iteration
(2) Z? H = upper bound for 1 -second hash iteration
(3) Si = security level (Z = 1， 2， 3， . . .， x), where x = 20， 32， or other values
(4) Si = nighest security level, s x = lowest security level
V
Determination of bound bi for each security level め is as follows:
(1) bi ᅳ— 0.2b L + 2 8 x (/ - 1), bi < 2. A
(2) i^i-l
(3) If i = 0， exit; else go to step 1 of Box 1402 again.
Figure 7.1 Pseudo-code to determine the numbers of hash iteration for multiple
security levels of multihash key methods and systems
119
Figure 7.2 presents the basic model of multihash key to generate multihashes
as site keys. A user needs to remember the selected security level for a specific
account. In case of forgetfulness, all the 20 security levels shall be tried one by one.
1500
V
Settings to create various slave keys d s (aka site keys) of multihash key:
(1) Necessary entries: Master key d, numeric y-digit passcode d n where y can be 4
(2) Optional entries: Username ID, domain name URL, or else NULL
(3) Bounds of hash iteration for various security levels s, : b u b 2 , b 3 , b b . . .， b x
(4) User selects security level S} among x security levels, where x = 32 or others
(5) Use 2n-bit hash function where 2n > 512 like SHA-512
Processing the master key d and passcode d n to create the determinants H b of hash
iteration number for each security level within their bounds:
(1) H b <— SHA-512 (d II d n ， 1) for one round of hash iteration
(2) H b (z\ ， Zi) = bit truncation of H b from bit z\ to bit zi
V
Calculate the hash iteration number j of a slave key:
(1) Choose either Fixed or Random j
(2) if Fixed
if i =1， j ᅳ (ゎ广 2 8 +l) + H b (0 , 7), j<b, ；
else if 1 < i<xj (ᄊ― i+l) + H b (8i—8 , 8/-1), j < b t ；
else if i = x， j (b x _i+l) + 仏 (&c— 8 , 8x— 1)， j <b x .
else if Random
j <— random[b i-2 H -\-l ， わ J ， where human remembers a random value
Generate slave key d s :
(1) Do if ID = URL = NULL, ^ SHA-512(J , j) ;
else if ID = NULL, Hi SHA-512(J II URL ， f) ；
else if URL = NULL, Hi SHA-512(J IIID,j) ；
else if ID and URL are not NULL, Hi SHA-512(J II ID II URL J).
(2) H Hi(0 ， 255)， n-bit truncation of Hi from MSB bit, where n = 256
(3) d s ᅳ— Bin2Txt{ n-bit CSPRBG (H) )， Bin2Txt = Binary-to-text encoding
N レ
Apply the slave key d s . Then, clear the memory storing all forms
of secrets and close all the application software.
Figure 7.2 Operation of the basic model of multihash key method and system
120
예 505
Necessary entries are master key d and numeric 4-digit passcode d^- Optional
entries are username ID and website (or domain name) URL. The username and
website are used to create diversity of multihash key from a key pair (master key,
passcode). Domain name can also help to resist phishing and spoofing attacks.
The 512-bit hash of the concatenated master key and passcode is truncated
into 20 partitions with 8-bit each from the MSB bit. This increases the randomness of
specific keys for different accounts. If an attacker does not know the exact security,
then 5120 hashes have to be checked for any key pair (master key, passcode). If the
attacker knows about the security level, then 28 hashes have to be validated for any
key pair (master key, passcode).
For the settings of bound &/， it can be either fixed or random. If the fixed
option is chosen, the number or hash iterations will use the standard settings. A user
is mobile and can use this method without remembering the number of hash
iterations while accessing offline and online login account from different computing
systems.
If the random option is chosen, the number of hash iterations will be
randomly selected by a user within a given range. User's mobility is weakened
unless one can remember the random values of hash iterations while accessing
offline and online logins. However, if a user can remember the hash iterations, this
option offers stronger resistance to dictionary attack. The best option is a hybrid
scheme. Choose fixed option for lower security levels and random option for higher
security levels.
Depending on the value existence of username ID and domain URL, the
master key undergoes different key hashing and key strengthening using SHA-512 to
generate hash Hi. Hi is then encoded from Dinary to text to fulfil the demands of
password requirements such as alphanumeric, mixed lowercase and uppercase, and
with punctuation marks.
Here, a binary-to-text encoding of Bin2Txt(H) is proposed as in Table 7.1.
Base64 encoding is not used as there are only two punctuation marks included
(Borenstein & Freed, 1992). Bin2Txt(H) converts 6 binary bits into one 8-bit ASCII
121
character. It has a bit expansion of 33%. All types of ASCII characters are included:
lowercase, uppercase, digit, and punctuation marks. The last group of 4 binary bits of
H from 253rd to 256th is padded with 2 binary bits of at the right or LSB side. The
output of Bin2Txt(H) is a string of 43 ASCII characters and is used as key hash.
Table 7.1 Binary-to-text encoding Bin2Txt(H) of multihash key
Bin
00
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
Txt
a
b
c
d
e
f
g
h
i
j
k
1
m
n
P
Bin
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Txt
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Bin
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Txt
1
2
3
4
5
6
7
8
9
<
>
？
Bin
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
Txt
I
#
$
%
&
(
)
*
+
1
@
Bin
Padding
Txt
N.B.: Bin: For easy understanding, decimal value is shown to represent Dinary values
Lastly, copy the hash as site key into the clipboard and paste it on the prompt
key field for authentication access. Remember to clear the clipboard before
leaving the computer.
On how to change from an old key into a new one, a user can change either
the master key, passcode, security level, username, or the domain name. There are
also proposed usages of 20 security levels as shown in Figure 7.3.
New methods and systems called multihash key and its variants are presented
here to generate multiple slave keys (aka site keys) from a single master key for both
the offline and online accounts. Among various cryptographic, information- hiding,
and non-cryptographic applications needing secrets for various types of key, here are
122
some of the popular applications of secret key: (i) Master key for password vault
hiding various keys; (ii) Internet banking; (iii) online stock trading; (iv) insurance; (v)
tax; (vi) office, school and home email accounts; (vii) instant messengers; (viii)
encrypted files; (ix) database accounts at the office and school; (x) library accounts;
and (xi) verification key for credit card. Hence, the impact contribution of multihash
key shall be very high in the aspects of reducing the human memorization burden
and system operating costs.
The multihash key method and system uses the hash iteration and hash
truncation, followed by optional /i-bit CSPRBG to increase the randomness, as for a
basic model of multihash key as in Figure 7.2， to generate slaves keys from a master
key and an optional passcode. The master key and hash function shall be at least 2n
bits. The passcode shall be at least 4 digits or more. The hash iteration applies the
key strengthening for a period ranging from 0.2 to 2 seconds, or longer to 10 seconds
in some of the variants of multihash key. Hash truncation halves the hash value or
message digest. Multihash key supports infinite number of online accounts and
limited number of offline accounts depending on the performance of the computer.
Examples of online accounts are webmail, login, email, and instant messenger.
Examples of offline accounts are encrypted file, public-key certificate, bank ATM
card, and software token.
Security levels: usages
1 Password file and key management tool like password vault.
2 Finance=>Very important Internet banking.
3 Finance=>Important Internet banking.
4 Finance=>Stock trading.
5 Finance=>Insurance， income tax, ...
6 Very important personal encrypted files, email accounts, instant messengers.
7 Important personal encrypted files, email accounts, instant messengers, ...
8 Very important accounts in working/ studying place like email.
9 Important accounts in working/ studying place like database.
10 Other accounts in working/ studying place like library.
1 1 〜； 20 Other not frequently used offline and online accounts.
Figure 7.3 Proposed usages of 20 security levels
123
Table 7.2 Comparisons of key management tools
Feature s\Key management tools
Plain
browser
Password Password Windows
LPWA
HP site
CPG
Password
SPP
Pwd
Hash
Passpet
Multihash
key
saie
live i ᄂ/
password
multiplier
Usability
1. Make logging in more convenient
No
Yes
No
Yes
Yes
No
？
Yes
？
No
Yes
？
2. Work with existing websites
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
3. Allow site -by -site migration to
tool
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
4. Change individual site keys
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
5. Log in from other computers
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
6. Only need to memorize one secret
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
7. Enable changing the master secret
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
8. Applicability to offline accounts
Yes
No
No
No
No
No
No
No
No
Yes
9. Applicability to online accounts
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
10. Integrate usages together with
other tools
Yes
No
No
Yes
No
No
No
No
No
Yes
Security
1 ᅳ Unique key for each account
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
2. Resist offline dictionary attacks
No
No
No
No
No
No
No
Yes
No
No
Yes
Yes
3. Adapt to increasing CPU power
No
No
No
No
No
No
No
Yes
No
No
Yes
Yes
4. Avoid storing keys
Yes
No
No
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
5. Avoid a single central authority
Yes
Yes
Yes
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
o. Resist phi shin g by fake login
forms
No
No
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
7. Resist mimicry of browser UI
No
No
No
No
No
No
No
No
No
No
Yes
No
8. Help the user identify websites
No
No
No
No
No
No
No
No
No
No
Yes
No
9. Stop entering secrets in webpages
No
No
No
No
No
Yes
？
Yes
？
No
Yes
？
Possible implementation
1 ᅳ Stand-alone application
No
No
Yes
No
No
Yes
No
Yes
Yes
Yes
No
Yes
2. Single sign-on server
No
No
No
Yes
Yes
No
Yes
No
No
No
No
No
3 ᅳ Browser extension
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
？: Unknown situation depending on implementation
7.2.4 Discussions
Table 7.2 compares various key management tools with multihash key from
the aspects of usability, security, and possible implementation. A lot of comparisons
are attributed by Yee and Sitaker (2006) on Passpet. New features used for
comparisons are applicability to offline and online accounts, integrated usages
together with other key management tools and possible implementations. It is
124
important to note here that multihash key can be used together with Passpet to earn
"Yes" for items [L7-L9] under the security features in Table 7.2.
Multihash key can be used for both offline and online accounts. Possible
implementations are stand-alone application and browser extension. These are simple
interfaces to input a password or key with unique key images for multiple accounts.
Memorizability is improved since there is only one secret for various login accounts.
Server is not used and hence there is no central authority. There are no single
point of failure and high cost of integration. It is mobile and there is no encrypted
storage of site keys. Since there is no integration, multihash key can be used for any
existing computer systems.
The passcode is optional to be remembered by a user because it can be
converted to be an 8-bit password supplement in one of the two methods of key
strengthening (Manber, 1996; Abadi, Lomas & Needham, 1997， 2000). Master key is
the password, and when it is combined with the password supplement, they form the
full password. Another key strengthening method is also called key stretching, which
uses a large amount of hash iterations (Kelsey, Schneier, Hall & Wagner, 1997).
The variant of SHA-2, which is SHA-512, is used in the key hashing and key
strengthening. This is because there are possible collision attacks to MD5 and SHA-1.
The hash truncation creates a 256-bit hash as site key. The unused truncated bit
creates a 128-bit security strength (256/2=128) preventing the compromised site keys
at the higher security level from revealing the site keys at the lower security level.
The passcode also has this feature but is very much less powerful.
For the experimental data of lower bound b L and upper bound bu of some
computer systems, there are reported as follows. For instance, for the first computer
system of desktop PC [Example 1: Pentium II 266MHz， 192MB RAM, running on
Windows XP Professional Edition], the lower and upper bounds for 1 -second hash
iteration, as in Figure 7.1, are 7600 and 8200， respectively. In other words, the first
computer system can only support 20 offline accounts for a security level
partitioning of 8 bits or 2 8 .
Yet in the second computer system of laptop PC [Example 2: Centrino Duo
1.66GHz, 1.5GB RAM, running on Windows XP Home Edition], the lower and
125
upper bounds for 1 -second hash iteration are 81,700 and 93,700 respectively. For this
specification, the second computer system can support 256 offline accounts for a
security level partitioning of 8 bits or 2 8 .
Yet in the third computer system of desktop PC [Example 3: Pentium D
2.80GHz, 512MB RAM, running on Windows XP Professional Edition SP3]， the
lower and upper bounds for 1 -second hash iteration are 52,500 and 122,500
respectively. For this specification, the third computer system can support 256 offline
accounts for a security level partitioning of 8 bits or 2 8 .
Using the proposed settings, the key strengthening has an access time from
0.2 second to 2 seconds. This is an efficient range of acceptable login processing
time. It can be calibrated to be parallel with the advances in computer technologies
for new releases of multihash key. Moore's Law is a good rule to judge the
calibration, which is about one bit faster for every two years (Wikipedia, 2008v).
Key hashing and key strengthening are also good techniques to resist offline
and online dictionary attacks as well as pre-computation attacks. To prevent phishing
and spoofing attacks, multihash key can either be used together with other anti-
phising tool like petname system and Passpet, or include domain name URL in its
key hashing. Malicious server attack is also prevented as different accounts have
unique passwords. For homograph attack due to visually similar Unicode graphic
symbols, the implementation of multihash key shall support the Unicode characters.
Up to here, the basic model of multihash key can support infinite number of
online account. Meanwhile, the number of supported offline account by multihash
key is given by Equation (7.1). From Figure 7.2， the security level x can be increased
up to the maximum of hash iteration number ゾ max . Also, hash functions beyond 512
bits like 768 and 1024 bits may be needed.
Saco = X (7.1)
7.2.5 Limitations
Multihash key can be implemented as a stand-alone application with no
change of setting at the server side. However, it is vulnerable to password file
126
compromise attacks and message log file attacks. Nevertheless, domino effect of
password reuse can be avoided. To get rid of password file compromise attacks and
message log file attacks, some countermeasures (Gouda, Liu, Leung & Alam, 2005)
can be adopted by changing the settings of authentication approach at client and
server sides.
Acting as a stand-alone application, multihash key requires a user to perform
extra steps. These steps are creating a key, copying, and pasting it to a login
prompted textbox. The user also needs to remember the security level of an account,
an at least 128-bit master key, and a numeric 4-digit passcode. These cause the
solution to be not user-rnendly.
To facilitate the application, multihash key has to be integrated into the user
interface of each authentication application. Therefore, the item [LI] of usability in
Table 7.2 about convenient logging in depends on implementation.
For security level, it can be jotted into a notebook in plaintext form because it
is not an essential secret. Alternatively, for online account, the user can be reminded
about the security level whenever the user sends the username to the server. This
allows an attacker to reduce the number of hash testing by 20 times, or 4.32 bits
(=log 2 20).
For numeric 4-digit passcode, it gives an extra security of 13.29 bits
(=log 2 10 000) and is not an essential secret. This passcode can be made constant for
user with poor memory. For user who can remember 128-bit master key, 4-digit
passcode and security level, the effective security strength is 145.01 bits. For user
who can remember only the 128-bit master key, its security strength is 128 bits.
Hence, security can be compensated for better usability.
Using multihash key, limited number of multiple offline and online accounts
can be supported as compared to the almost infinite number of online accounts for
LPWA, HP Site Password, CPG, Password Multi-plier, SPP, PwdHash, and Passpet.
For more accounts, faster computer system is needed to have larger bound range. Or
else, the partition between any two security levels has to be reduced.
127
7.2.6 Conclusion
The proposed invention of multihash key requires users to remember a master
key and passcode to generate unique key hashes or site keys for multiple accounts.
For security level, username, and domain name of a specific account, users can
choose to write them down somewhere as there are not critical secrets. This is a
balance between the usability and security.
Multihash key can be used for offline and online accounts, where existing
similar key management tools without encrypted site key storage can only be applied
to online accounts. It is hoped that this proposal can release the human memory
burden on required passwords or keys for various types of increasing accounts. To
have better resistance to phishing and spoofing attacks, try to use multihash key
together with an anti-phishing tool like petname system and Passpet.
7.3 More Slave Keys for Offline Accounts per Master Key of Multihash Key
Key hashing of master key and domain name generates site keys for online
account only like login. Meanwhile, multihash key using hash truncation can be
applied for both online account and offline account like password encryption of
electronic file using symmetric key cryptography and partial memorizable private
key of the asymmetric key cryptography. Nevertheless, even though infinite online
account can be supported by multihash key, only limited offline account is possible.
The current computer speed limits it to 20 offline accounts unless we have faster
computer system, reduced partition length, or more passcodes.
Here, three methods are proposed to support more offline accounts per master
key of multihash key using filename, random number, or two -tier structure.
Multihash key using a random number can now have up to 2 256 site keys for
electronic file of offline account. Meanwhile, multihash key using a two -tier
structure can support multiple partial memorizable private keys of the split private
key cryptography up to 400 offline site keys.
To support more offline accounts, especially the various cryptographic
schemes of MePKC, multihash key is further enhanced. Firstly, hashing the
128
concatenation of a master key and a filename is proposed as in Figure 7.4a. As long
as the filename is unique, infinite offline accounts can be supported. However, the
problem is name clashing and renaming.
Secondly, a random number is used without and with multihash key,
respectively, as in Figures 7.4b-7.4c， where this random number is concatenated with
master key in a hashing process to generate a slave key. For a ciphertext encrypted
using this slave key, the random number has to be retrieved first. Hence, this random
number is encrypted using the master key and stored as a concatenation to a file
ciphertext encrypted by the slave key to become an output file. When a user wants to
open the file ciphertext, one splits the output file to get the ciphertexts of file and
random number. Decrypt the ciphertext of random number using the master key.
Then, generate the slave key using the master key and the recovered random number.
Subsequently, the file ciphertext is decrypted by the slave key. Using AES-256, this
method using a random number can support 2 256 offline accounts. However, its
drawbacks are major modification to the current computer systems and no support
for secrets of offline accounts without any ciphertext storage, such as split private
key crypto system and MePKC.
Then, a third method, as in Figure 7.4d, using a two-tier structure of
multihash key is proposed. For the examples of the first and second computer
systems, 400 and 65536 offline accounts, respectively, can be supported. This
method is compatible with the current computer system. Yet the special advantage of
this method is its support for secrets of offline accounts without any ciphertext
storage. In other words, the partially and fully memorizable private keys of split
private key crypto system and MePKC can now be supported.
1600
Using Filename
P = plaintext, Km = master key, F
= file name, Ks = slave key, Ci = file ciphertext,
Encryption
Decryption
K s Hash (K M II F)
K s Hash (K M II F)
Ci ― Encrypt (P ， Ks)
P Decrypt (Ci ， Ks)
Figure 7.4a rirst method: Using filename
129
1601
Using Random Number without Multihash Key
P = plaintext, Km = master key, R = random number, Ks = slave key,
Cr = R ciphertext, Ci = file ciphertext, C2
= output file,
Truncation = n-bit truncation, j = hash iteration number
Encryption
Decryption
K s ― Truncation ( Hash (K M II R) )
[C R ， Ci] ᅳ Truncation(C 2 )
Ci ᅳ Encrypt (P ， Ks)
R Decrypt (C R ， K M )
Cr <— Encrypt (R ， K M )
K s Truncation ( Hash (K M II R) )
C 2 ᅳ (Cr II CO
P Decrypt (Q ， Ks)
Figure 7.4b Second method: Using random number without multihash key
1602
Using Random Number with Multihash Key
Encryption
Decryption
K s ― Truncation ( Hash (K M II R J) )
[C R ， Ci] ᅳ Truncation (C 2 )
Ci ― Encrypt (P, Ks)
R Decrypt (C R ， K M )
Cr <— Encrypt (R， K M )
K s Truncation ( Hash (K M II R J) )
C 2 ― (Cr II d)
P Decrypt (Ci , Ks)
Figure 7.4c Second method: Using random number with multihash key
1603
Using Two-Tier Structure of Multihash Key
Settings: Multihash = multihash key function, Km = master key, Kn =
passcode K S i = first slave key, Ks 2 = second slave key, j\ = first hash
iteration number, 72 = second hash iteration number, Ks = final slave key
for various offline accounts
(1) Ksi ᅳ Multihash ( K M ， K N ，力 )
(2) Ks2 ᅳ Multihash ( Ksi , K N J 2 )
(3) Stored in notebook (j\ ,j 2 )
(4) K s ― K S2
Figure 7.4d Third method: Using two-tier structure of multihash key with manually
selected security levels
Figure 7.4 Methods and systems to support more offline accounts for multihash key
130
7.3.1 Related Works
In the information era, human has many keys or passwords for various offline
and online accounts. On 4 February 2001， Kanaley (2001) with Forrester Research
reported that an active Internet user averagely manages 15 keys on a daily basis. On
the other hand, Adams and Sasse (1999) reported that a user can only be expected to
cope with 4 to 5 keys that were unrelated and regularly used. For user's unique keys
without the constraint of relevancy, Florencio and Herley's survey (2007) reported
an average 6.5 keys, repeated 3.9 times each for 25 accounts and typing 8 keys daily.
This reflects a demand to enhance human memorizability to support many accounts
needing a key. Online account is like email and webpage login. Offline account is
like file encryption, encrypted private key, and memorizable private key of MePKC.
Key hashing of master key and domain name (or server name) is used to
generate site keys for online accounts (Gabber, Gibbons, Matias & Mayer, 1997;
Karp, 2003; Karp & Poe， 2004; Halderman, Waters & Felten, 2005; Gouda, Liu,
Leung & Alam, 2005; Ross, Jackson, Miyake, Boneh & Mitchell, 2005; Yee &
Sitaker, 2006). Other methods of single master key for multiple online site keys are
single sign-on server (Windows Live ID, No date) and CPG (Common Password
Method) using website account with assigned random number (Luo & Henry, 2003).
Meanwhile, multihash key (Lee & Ewe, 2007a) using key strengthening and
hash truncation can be applied for both online account and offline account.
Nevertheless, even though infinite online account can be supported by multihash key,
only limited offline account is possible. The current computer performance limits it
to 20 offline accounts unless the computer system is faster or reduced the partition
length between the two security levels.
However, the computer speed increases by 1 bit for every 24 months. It is not
surprising where one may have a few hundreds of accounts up to 512. We may need
to wait for 10 years if computer speed is expected to increase the supported offline
account. For the reduction of partition length, it decreases the attacking time because
the security level of those account holders with poor memory is public. The current
partition length at 2 8 is already considered short. It is not encouraged to be shorter.
131
Instead, the better approach is to increase the partition length in order to increase the
attacking time whenever the security level is known.
Hence, the only current possible practical way to have over 20 offline
accounts is to change the passcode (Lee & Ewe, 2007a). However, this method is not
appropriate for those with poor memory. Moreover, the increment of offline account
is additive, which means an increment of 20 offline accounts for every additional
passcode. Here, three methods for multihash key are proposed to support more
offline accounts using filename, random number, or two -tier structure.
7.3.2 First Method: Using Filename
Domain name or server name is commonly used to generate infinite slave
keys (aka site keys) for online ac-counts from a master key (Gabber, Gibbons, Matias
& Mayer, 1997; Karp, 2003; Karp & Poe， 2004; Halderman, Waters & Felten, 2005;
Gouda, Liu, Leung & Alam, 2005; Ross, Jackson, Miyake, Boneh & Mitchell, 2005;
Yee & Sitaker, 2006). The general equation is in Equation (7.2).
Slave Key = Hash (Master Key II Domain Name) (7.2)
For offline account like file encryption and encrypted private key, a tagged
filename informs that no change can be made. Then, Equation (7.3) can be used to
support over 20 offline accounts, where filename replaces the domain name. This
method using filename is shown in Figure 7.4a.
Slave Key = Hash (Master Key II Filename) (7.3)
As long as the filename is different, infinite offline account can be supported.
The disadvantage of this method is that the computer file system has to be
restructured so that there is an option to ban the change of filename. Having another
file to hint that a specific filename of a file cannot be changed may have problem
because a user may forget to transfer and store these two files together.
132
Another problem is that due to different computer operating systems, the
filename with some special characters like space mark and some punctuation marks
tends to change automatically when it is transferred and saved from one computer to
another computer.
Thirdly, for non- ASCII multilingual filename like CJK filename, it may have
problem as well when the file is transferred and saved among different computing
systems. For encrypted private key in the form of smartcard, there exists the problem
that no filename is available. Consequently, this method using filename can only be
conditionally implemented independent of or together with multihash key.
7.3.3 Second Method: Using a Random Number
CPG (Luo & Henry, 2003) assigned a random number for every website
account and storea it in a server to generate unique site keys for different online
accounts. It is not designed for offline account at all. Here, a new algorithm is
designed using random number as well to support over 20 offline accounts up to the
set size of the random number. This new algorithm can be implemented independent
of or together with multihash key. Figure 7.4b shows an independent implementation
from multihash key; whereas Figure 7.4c shows an implementation together with
multihash key.
Let P = file plaintext, R = random number, Km = master key, Ks = slave key,
Cr = ciphertext of R, Ci = file ciphertext, C2 = output, Hash(X II Y) = hashing the
concatenation of X and Y for one round, Hash(X \\Y ,b) = hashing the concatenation
of X and Y for b round, Enc = encryption, Dec = decryption, Trunc = truncation, II =
concatenation, and B(x) = bit length of x. Equation (7.4) is the setting.
B(K S ) = B(K M ), B(Hash) > 3B(K M )， B(R) < B(Hash) - B(K M ) (7.4)
For security analysis of both the algorithms using random number in Figures
7.4b-7.4c, an attacker will try to recover the random number R and file plaintext P.
The ciphertext of random number Cr cannot be cryptanalyzed as long as it is random,
which means there is no loophole in the pseudorandom number generator (PRNG).
133
Another condition for the random number R is that its bit length is maximized as in
Equation (7.3) with promising random number generation speed.
When the cryptanalysis of Cr is impossible, the only attacking place is the
cryptanalysis of file ciphertext Ci. Ci can be obtained easily because the truncation
function of C2 to get Cr and Ci is public. For algorithm in Figure 7.4b independent
of multihash key, guessing attack and pre-computation attack are possible and take
less computing efforts than the algorithm in Figure 7.4c. This is because the pre-
computation attack over the algorithm in Figure 7.4b requires only one round of
hashing to generate the slave key Ks. Meanwhile, algorithm with multihash key in
Figure 7.4c requires many rounds of hashing. The extra security strength of Figure
7.4c algorithm from Figure 7.4b algorithm is because of the key strengthening
(Manber, 1996; Abadi, Lomas & Needham, 1997， 2000; Kelsey, Schneier, Hall &
Wagner, 1997). Key strengthening, which is also called key stretching, is a process to
hash an input value for many rounds until a response time of about 1 second.
The hash truncation in both Figures 7.4b-7.4c introduces more difficulty to
get master key Km from compromised Ks. If an attacker can recover P and Ks from
Cr, it is still very hard for him to recover Km. Whenever the Km is not compromised,
the compromise of a single offline account will not affect the other offline accounts.
Of course, the security of the hash function is also very important. The security
strength of a hash function is half of its bit length. In case of compromised Ks,
B(Hash) has to be at least triple of B(K M ) as in Equation (7.3). For instance, if SHA-
384 is used, the K M ， Ks, and system security strength are all 128 bits. For hash
function longer than 512 bits, a developer can refer to scalable polymorphic hash
(Roellgen, No date, 2005).
For computation load of algorithm in Figure 7.4b, it runs faster than
algorithm in Figure 7.4c. However tms faster speed is not good because it also means
faster possiole cryptanalysis. Figure 7.4c algorithm runs slower but it has acceptable
response time with key strengthening set at about 1 second. The delay of response
time is significant as it also freezes the requirement of longer key size as the
computer technologies are advancing. The number of hash iteration can be simply
increased to maintain a response time at about 1 second. For random number
134
generation, Le Quere (2004) invented a fast RNG with 100 Mbps. Its drawback is
that a special physical RNG is needed. Another alternative without special hardware
is a PRNG based on discrete logarithm problem by Gennaro (2005) with 860 bits per
160-bit exponent exponentiation when typical parameters are n = B(modulo prime) =
1024 and c = B(exponent) = 160.
The first weakness of this method using random number is that it cannot be
implemented on the current encryption system. Only new encryption system with
updated software program and hardware device can adopt it. Also, for some special
cases where there is no ciphertext, the ciphertext of a random number has no place to
be stored. This happens to the private key storage technology like split private key
(Sandhu, deSa & Ganesan, 2005c, 2006c, 2006f). Different asymmetric key pairs
may be needed for different purposes like encryption, digital signature, key exchange,
authentication, digital timestamping, and multipartite public key cryptography for
electronic commerce.
7.3.4 Third Method: Using a Two-Tier Structure
The third method is proposed to get rid of the disadvantages of the second
method. It can only be used together with multihash key. For the mean time, the
increment of supported offline account is additive with extra 20 site keys for every
additional passcode. Here, the increment of the slave keys for offline accounts can be
made to be multiplicative by having a two-tier structure.
For this two -tier multihash key, there are two rounds of multihash key. The
slave key from the first round of the multihash key (intermediate site key) will
become the master key of the second round of the multihash key (intermediate
master key). The slave key from the second round of the multihash key (final slave
key) will be used for the offline account like rile encryption and private key.
Subsequently, there are 20*20 = 400 site keys for offline accounts. This will
be sufficient for the current usage demand of asymmetric key pair. In case it is not
enough, the two-tier structure can become three-tier structure or higher order. In case
it is a three-tier multihash key, there are 20*20*20 = 8000 site keys for offline
accounts. Figure 7.4d shows a two -tier multihash key.
135
For this method, there exists a problem to remember the combination of
security level. A user may need to save this combination somewhere in a paper
notebook or encrypted electronic file. Consequently, portability is a weakness in case
the notebook is not around. However, its advantages are support to old encryption
system and applicability to offline account without ciphertext storage like the
memorizable partial private key of split private key technology (Sandhu, deSa &
Ganesan, 2005c, 2006c, 2006f) and fully memorizable private key of MePKC.
7.3.5 Conclusion
Three methods to increase the number of slave keys for offline accounts have
been proposed: Using a filename, random number, and two -tier structure. The first
method is the simplest, but has impractical implementation conditions. The second
method is good for new file encryption system with ciphertext storage.
Complementanly with the second method, the third method can be used for old and
new file encryption system without ciphertext storage. In a nutshell, human needs
memorizability enhancement like multihash key to have various slave keys from a
master key for both the offline and online accounts.
The software implementations (Lee, 2007a, 2008i， 2008k) of the multihash
key have been built up by using the Microsoft Visual Studio (Marshall, 2003) and
can be evaluated by downloading a copy of this software from [URL:
www.xpreeli.com].
7.4 First Variant of Multihash Key Using Automatically Selected Tiers and
Security Levels
Besides the basic model, multihash key has been innovated to have some
variants. The first variant in Figure 7.5 supports more offline accounts by using
automatically selected tiers and security levels.
Box 1701 gives the settings to create various slave keys d s (aka site keys) of
multihash key. Necessary entries are master key d, numeric y-digit passcode d n ,
where y can be 4， and sequence ID Q.
136
1700
V
Settings to create various slave keys d s (aka site keys) of multihash key:
(1) Necessary entries: Master key d, numeric y-digit passcode d n , sequence ID Q
(2) Optional entries: Username ID, domain name URL, or else NULL
(3) Bounds of hash iteration for various security levels s,- : b u b 2 , b 3 , b!, . . .， b x
(4) (d II d n II Q) selects security level Si among x security levels, where x = 32
(5) Use 2/i-bit hash function where 2n> 512 like SHA-512
(6) H h (z.\ ， z?) = bit truncation of H h from bit zi to bit z?
1701
Processing keys d, d m and Q to create determinants H b of hash iteration number j t
within their bounds and security levels i = x t for each tier of multihash key, and
then calculate the hash iteration number j t and security level x t of each tier t :
(1) for ? = to m， where m = 9 or other values,
H b <— SHA-512 (d II d n \\Q, 1) for one round of hash iteration ；
x t // fo (8x+?log 2 x ， 8x+ (？ +l)log 2 ^;-l) = H b {^x+5t ， 8x+5?+4) ；
iH=l， h-
else if 1 < j < x, j,
else h i = x, j,
ゅ广2 8 +1) + / も (0,7), j t < b x ；
(fon+l) + H b (8i—8 , 8/-1), j t < h ；
+ H b {%x-% , 8x-l), j, < b x ；
Generate intermediate slave keys H t for tier t :
if ID = URL = NULL, H it SHA-5 1 2(d ， j t ) ；
else if ID = NULL, H it SHA-5 12(J II URL ， j t ) ；
else if URL = NULL, H lt SHA-5 12(J II ID ， j t ) ；
else if ID and URL are not NULL, H it SHA-5 12(J II ID II URL J t ) ；
d n ^H t ^ H it (0 ， 255)， n-bit truncation of H it from MSB bit ； 1;
Go to step (1) of Box 1702 whenever t<m. end
Generate slave key d s :
(1) User selects H 아 H m or H 아 SHA-5 12(// ᅵᅵ^ i II 표 II .. ᅳᅵᅵ 하 II ... II H m )
(2) d s Bin2Txt(n-bit CSPRBG (H(0 ， 255)))， Bin2Txt = Binary-to-text encoding
/ 싀
1703
Jot down Q or store Q at a local/remote server for future access.
Apply the slave key d s . Then, clear the memory storing all forms
of secrets and close all the application software.
1704
Figure 7.5 First variant of multihash key method and system to support more offline
accounts using automatically selected tiers and security levels
137
Sequence ID Q can be in plaintext and is used to create multiple unique
offline and online slave keys. Q can be jotted down into a notebook, or stored at local
and remote servers for future acknowledgment to the user about the Q value of one's
account. Optional entries are username ID, domain name URL, or else NULL.
Bounds of hash iteration for various security levels s, are b u b 2 , & 3， ...， H ...， b x .
Concatenation of (d II d n II Q) selects security level Si among x security levels, where
X = 20， 32 or others. This method uses 2/i-bit hash function, where 2/7 > 512 like
SHA-512. H b (z\ ， zi) means bit truncation of H b from bit z\ to bit zi.
At Box 1702, master key d, passcode d n , and sequence ID Q are processed to
create the determinants Hb of hash iteration number j t within their bounds and
security levels i = x t for each tier of multihash key, and then calculate the hash
iteration number 요 and security level x t of each tier t. Here, an intermediate slave key
H t is derived at each tier and replaces the d n . Repeat step (1) in Box 1702 whenever
the maximum number of tier m has not been reached.
At Box 1703, final slave key d s is generated by directly taking the slave key
at the final tier or hashing the concatenation of derived secrets from each tier. At Box
1704, jot down Q or store Q at a remote server as like salt for future access, apply the
slave key d s , clear the memory storing all forms of secrets, and then close all the
application software. The passcode here can be optionally replaced by a big
memorizable secret for more randomness to support more offline accounts up to
Equation (7.5). Security level x can be increased up to the maximum of hash iteration
number j max . Also, hash functions beyond 512 bits like 768 and 1024 bits may be
needed.
Saci =x (7.5)
7.5 Second Variant of Multihash Key Using Automatically Selected
Permutation Sequence of Security Levels
The second variant in Figure 7.6 also supports more offline accounts by using
automatically selected permutation sequence of security levels.
138
1800
V
Settings to create various slave keys d s (aka site keys) of multihash key:
(1) Necessary entries: Master key d, numeric y-digit passcode d n , sequence ID Q
(2) Optional entries: Username ID, domain name URL, or else NULL
(3) Bounds of hash iteration for various security levels s, : b u b 2 , b 3 , . . .， b x
(4) (d II d n II Q) selects security level s t among x security levels, where x = 32
(5) Use 2n-bit hash function where 2n> 512 like SHA-512
V
Processing keys d, d m and Q to create determinants H b of hash iteration number j,-
within their bounds and permutation number pq (= p q ) to select a security level i :
(1) H b SHA-512 (d II 4 II Q , 1) for one round of hash iteration
(2) H b (z\ ， Zi) = bit truncation of H b from bit zi to bit zi
V
Calculate the hash iteration number j,- for each security level i :
(1) for i = 1 tox，
if i = 1， ji 수— {b x -2\\) + H b (0 ， 7), j'- < b x ；
else if 1 < / < xjt ᅳ (^！ +1) + 仏 (8Z— 8 ， 8/-1), j t < b { ；
else if i = x， j,- <— (b x _i+l) + //わ (8 ぶ— 8 ， 8x-l), j,- < b x . end
V
Generate intermediate slave keys H} for i=\to x and then slave key d s :
(1) for i = 1 tox，
if ID = URL = NULL, H u SHA-5 1 2(d ， j ( ) ；
else if ID = NULL, H u SHA-512(J II URL ， j" ) ；
else if URL = NULL, H u SHA-512(J II ID Jd ；
else if ID and URL are not NULL, H u SHA-512(J II ID II URL Jd ；
Hi H u (0 ， 255)， n-bit truncation ofH u from MSB bit . end
(2) Generate the permutation number pq (= p q ) for some selected Hi :
for ひ = 1 to floor(n/log 2 x), where floor(n/log 2 x) = 51，
M 수 Pq ᅳᅳ ^(8x+(^-l)log 2 x, 8x+^log 2 x-l) = H b (^x+5q-5, 8 ズ +5^—1).
(3) H ᅳ SHA-512^! II H p2 II H p3 II ... II / 뇨 II . . . II H p5l )
(4) d s Bin2Txt(n-bit CSPRBG (H(0 ， 255)))， Bin2Txt = Binary-to-text encoding
V
Jot down Q or store Q at a local/remote server for future access.
Apply the slave key d s . Then, clear the memory storing all forms
of secrets and close all the application software.
Figure 7.6 Second variant of multihash key method and system to support more
offline accounts using automatically selected permutation sequence of security levels
139
세 802
씨 803
〜顯
씨 805
Box 1801 gives the settings to create various slave keys d s (aka site keys) of
multihash key. Necessary entries are master key d, numeric y-digit passcode d n ,
where y can be 4， and sequence ID Q. Sequence ID Q can be in plaintext and is used
to create multiple unique offline and online slave keys. Q can be jotted down into a
notebook, or stored at local and remote servers for future acknowledgment to the
user about the Q value of one's account. Optional entries are username ID, domain
name URL, or else NULL. Bounds of hash iteration for various security levels S/ are
b\, & 2， & 3， ...， In, ...， b x . Concatenation of (d II d n II Q) selects security level s- t among x
security levels, where x = 20， 32 or others. This method uses 2n-bit hash function,
where In > 512 like SHA-512.
At Box 1802， master key d, passcode d n , and sequence ID Q are processed to
create the determinants Hb of hash iteration number ji within their bounds and
permutation number pq (= p q ) to select a security level i. H b (zi ， zi) means bit
truncation of H b from bit z\ to bit zi. At Box 1803， calculate the hash iteration
number 7/ for each security level i.
At Box 1804， generate intermediate slave keys H L at each security level and
then slave key d s . To select H h permutation number p q is used. The final slave key is
the hashing of the concatenation of multiple Hi based on p q . There may be a special
permutation number meaning NULL value where no bitstream is concatenated. If all
the selected H t are NULL, then select another d n and repeat all the steps.
At Box 1805， jot down Q or store Q at a remote server as like salt for future
access, apply the slave key d s , clear the memory storing all forms of secrets, and then
close all the application software. Let T be the maximum number of concatenated Hi
based on p q . The passcode here can be optionally replaced by a big memorizable
secret for more randomness to support more offline accounts up to Equation (7.6).
Security level x can be increased up to the maximum of hash iteration number j max .
Also, hash functions beyond 512 bits like 768 and 1024 bits may be needed.
(7.6)
140
7.6 Third Variant of Multihash Key Using Hybrid Combination
The third variant in Figure 7.7 is a hybrid combination of the first and second
variants to support more offline accounts using a hybrid combination of
automatically selected tiers and security levels, and automatically selected
permutation sequence of security levels.
Firstly, do the operations in Box 1701. Then, at Box 1900， master key d,
passcode d n , and sequence ID Q are processed to create the determinants H b of hash
iteration number j t within their bounds, permutation number pq (= p q ) to select a
security level i, and security levels i for each tier t of multihash key. Here, calculate
the hash iteration number ji for each security level i at tier t. Generate first
intermediate slave keys Hu for / = 1 to x at tier t. Generate the permutation number
pq (= p q ) for some selected Hu at tier t. Generate second intermediate slave keys Hit
for tier t and replaces the d n . Repeat steps (1.0-1.4) in Box 1900 whenever the
maximum number of tier m has not been reached. There may be a special
permutation number meaning NULL value where no bitstream is concatenated. If all
the selected Hi are NULL, then select another d n and repeat all the steps.
At Box 1901， final slave key d s is generated by directly taking the slave key
at the final tier or hashing the concatenation of derived secrets from each tier. At Box
1902， jot down Q or store Q at a remote server as like salt for future access, apply the
slave key d s , clear the memory storing all forms of secrets, and then close all the
application software. Sequence ID Q can be in plaintext and is used to create
multiple unique offline and online slave keys. Q can be jotted down into a notebook,
or stored at local and remote servers for future acknowledgment to the user about the
Q value of one's account. Let T be the maximum number of concatenated Hu based
on pq. The passcode here can be optionally replaced by a big memorizable secret for
more randomness to support more offline accounts up to Equation (7.7). Security
level X can be increased up to the maximum of hash iteration number j max . Also, hash
functions beyond 512 bits like 768 and 1024 bits may be needed.
(y=T \ m
S A C3= Z x> (7.7)
V y= l J
141
From 1701
V
Processing keys d, d m and sequence ID Q to create determinants H b of hash
iteration number j,- within their bounds, permutation number pq (= p q ) to select a
security level i, and security levels i for each tier t of multihash key :
(1.0) for ? = to m, where m = 9 or other values,
H b SHA-512 (d II d n \\Q , 1) for one round of hash iteration ；
(1.1) Calculate the hasn iteration number j,- for each security level i at tier t :
for Z = 1 to x,
if i = 1， ji {b x -2\\) + H b (0 ， 7)， j'- < b x ；
else if 1 < / < xji (씌 +1) + H b (8i—8 , 8/-1), j) < b } ；
else if i = x, j) <— (b x _i+l) + //わ (8x— 8 ， 8x— 1)， j,- < b x . end
(1.2) Generate first intermediate slave keys H u for / = 1 to x at tier t :
for Z = 1 to x,
if ID = URL = NULL, H Ui SHA-512(J , j" ) ；
else if ID = NULL, H Ui SHA-512(J II URL ) ；
else if URL = NULL, H Ui SHA-512(J II ID ， j ( ) ；
else if ID and URL are not NULL, H Ui SHA-512(J II ID II URL )
H u H Ui (0 ， 255)， n-bit truncation of H Ui from MSB bit . end
1900
(1.3) Generate the permutation number pq (= p q ) for some selected H u :
for な = 1 to floor(n/log 2 ぶ)， where floor(n/log 2 x) = 51，
^ p q ^ Hb(^x+(q-l)log2X, 8x+^log 2 x-l) = H^x+5q-5, 8 ズ +5^—1).
(1.4) Generate second intermediate slave keys H 2l for tier t :
d n ^H 2! iSHA-512(H lpl II H lp2 II H lp3 II ... II H lpq II ... \\ H lp5l )U ^ t+U
Go to step (1.0) of Box 1900 whenever t<m . end
Generate slave key d s :
(1) User selects H^H 2m orH^ SHA-5 12(// 20 \\H 2l \\ H 22 II ... \\H 2t \\ ... II H 2m )
(2) d s Bin2Txt(n-bit CSPRBG (H(0 ， 255)))， Bin2Txt = Binary-to-text encoding
N レ
Jot down Q or store Q at a local/remote server for future access.
Apply the slave key d s . Then, clear the memory storing all forms
of secrets and close all the application software.
Figure 7.7 Third variant of multihash key method and system to support more offline
accounts using a hybrid combination of automatically selected tiers and security
levels, and automatically selected permutation sequence of security levels
세 901
142
7.7 Fourth Variant of Multihash Key as a Further Authentication Factor
For the fourth variant in Figure 7.8， it is a specific application of multihash
key to act as a further authentication factor in the Internet banking, online share
trading, or other situations.
At Box 2001， bank and user apply a key exchange protocol to establish a
shared master key d, optional passcode d n , and initial downcount/upcount number N
for hash iteration in multihash key. Set N = N c initially. At Box 2002 for Internet
banking transaction needing a second authentication factor, it is triggered by a user
requesting for execution of a transaction that needs further authentication. Bank
server then sends a first message with random value R， timestamp T, current
downcount/upcount number N c to the remote user in a secure channel like SSL.
At Box 2003 for user response to the bank's challenge, user uses the
downcount/upcount number N c as the hash iteration number of a multihash key
process to generate a slave key d s \ from master key d and pin d n . Then, user uses the
slave key d s \ to encrypt the first message to create a second message using symmetric
key cipher. Later, user sends the second message as response to the bank server in a
secure channel like SSL for further authentication.
At Box 2004 for verification of user's response by bank server, bank uses the
downcount/upcount number N c as the hash iteration number of a multihash key
process to generate a slave key d S 2 from shared keys d and d n . Then, bank decrypts
the second message using slave key d S 2 to get a third message. If the first message
and third message are identical, then the user is verified and authenticated for further
user- selected transaction. Otherwise if the first message and third message are not
identical, then the user is rejected for further user- selected transaction. If the user is
verified for further authentication, decrement the N c by one unit for downcount, or
increment the N c by one unit for upcount. If the user is rejected for further
authentication, user chooses to go to step (1) in Box 2002 for re-try or go to Box
2005 for exit. For re-try or new request for further authentication, go to step (1) in
Box 2002. Otherwise, go to Box 2003 to clear the memory storing all forms of
secrets and close all the application software.
143
2000
\レ
Bank and user apply a key exchange protocol to establish a shared master key d,
optional passcode d m and initial downcount/upcount number N for hash iteration
in multihash key. Set N = N C initially.
\
f
Internet banking transaction needing a second authentication factor:
(1) User requests for execution of a transaction that needs further authentication
(2) Bank server sends a first message with random value R， timestamp T， current
downcount/upcount number N c to the remote user in a secure channel like SSL
、
f
User response to the bank's challenge:
(1) User uses the downcount/upcount number N c as the hash iteration number of a
multihash key process to generate a slave key d sX from master key d and pin d n
(2) User uses the slave key d s \ to encrypt the first message to create a second
message using symmetric key cipher
(3) User sends the second message as response to the bank server in a secure
channel like SSL for further authentication
V
Verification of user's response by bank server:
(1) Bank uses the downcount/upcount number N c as the hash iteration number of
a multihash key process to generate a slave key d s2 from shared keys d and d n
(2) Bank decrypts the second message using slave key d s2 to get a third message
(3) If the first message and third message are identical, then the user is verified
and authenticated for further user-selected transaction
(4) Otherwise if the first message and third message are not identical, then the
user is rejected for further user-selected transaction
(5) If the user is verified for further authentication, N c <— N c - I or N c <— N c + I
(6) If the user is rejected for further authentication, user chooses to go to step (1)
in Box 2002 for re-try or go to Box 2005 for exit
(7) Re-try or new request for further authentication?
No
Clear the memory storing all forms of secrets and close all the
application software.
2001
2002
2003
2004
/XT'
2005
Figure 7.8 Fourth variant of multihash key method and system for the specific
application to act as a further authentication factor in the Internet banking or other
situations
144
7.8 Fifth Variant of Multihash Key as a Simple Key Escrow Method and
System
The fifth variant in Figure 7.9 is another specific application of multihash key,
where it acts as a simple key escrow method and system for supervisor-wise non-
critical secrets.
Simple Key Escrow Method and System Using Key Management of Multihash
Key for Supervisor-wise Non-critical Secrets
Kgm = grandmaster key, SID = stafr identity number, Y = current year,
K S m = staff master key, Ks S = staff slave key, CID = client identity number,
EID = event identity number, Kcs = client slave key,
Multihash = multihash key function as in Figures 7.2， 7.5-7.7.
Generation of Staff Slave Keys Generation of Client Slave Keys
Supervisor holds key K GM Ks M Kss
Kss Multihash ( Kg M II SID II EID II Y) Kcs 아 Multihash ( Ks M II CID II EID II Y )
Both supervisor and staff know key Kss. Both staff and client know key Kcs.
Key Escrow
Slave keys and master keys at a lower key management levels are known to people
holding master keys and grandmaster keys, respectively, at a higher management
level.
Figure 7.9 Fifth variant of multihash key method and system for the specific
application to act as a simple key escrow method and system for supervisor-wise
non-critical secrets
Key management of multihash key is applied here. Slave keys and master
keys at a lower key management levels are known to people holding master keys and
grandmaster keys, respectively, at a higher management level. For the generation of
staff slave keys, a supervisor holding grandmaster key Kgm uses the stafr identity
number SID, event identity number EID, and current year Y, to generate staff slave
keys Kss from multihash key for different applications, where Kss Multihash
( Kgm II SID II EID II Y ).
A staff stores all one's staff slave keys into one's password vault. For the
generation of client slave keys, a staff slave key becomes a staff master key Ksm.
145
Ksm is used together with client identity number CID, event identity number EID,
and current year Y to generate client slave keys from multihash key again for
different applications, where Kcs Multihash ( K SM II CID II EID II Y ). A client
stores all one's client slave keys into one's password vault. In this way, the higher
management people have escrowed the slave keys at the lower levels. This approach
can be used for supervisor-wise non-critical secrets but confidential to the external
parties.
7.9 Discussions on Variants of Multihash Key
Variants 1， 2， and 3 optionally require the passcode to work automatically or
are upgraded to become a big memorizaole secret created as in Figure 3.1. After the
passcode has been replaced by a big memorizable secret with at least 128 bits, the
sequence ID Q can be optionally used to make the generated slave keys unique.
Yet in the current Internet banking, a random number in an SMS (Short
Message Service) through mobile phone network, or a one-time-password token
(OTP token), like RSA SecurlD token, is used as a second authentication factor.
Meanwhile, variant 4 alternatively uses downcounting or upcounting of hash
iteration number to generate various slave keys from a master key to function as the
second authentication factor. Lastly, variant 5 is designed for the key management of
supervisor-wise non-critical secret in an organization like government, company,
university, and school, to function as a simple key escrow method and system.
146
CHAPTER 8 MULTIHASH SIGNATURE
Since the introduction of classical digital signature scheme in 1976， many
variants are created, such as blind signature, multisignature, group-oriented signature,
threshold signature, etc. These variants perform certain specific functions of digital
signature. Here, a new variant called multihash signature (aka object-designated
signature) is proposed, where a single message can have multiple digital signatures
from a single asymmetric key pair. This is done so by having a few rounds of
hashing to produce multiple hashes, which are then signed by a private key to
generate multiple unique digital signatures. All of these digital signatures can be
verified using the same public key. This variant has some new message management
functions, which are to trace a file downloaded from different mirror sites, to referee
an advertiser broadcasting the news of a sponsor, and to monitor the leaking source
that publicly discloses a classified digital file such as will, contract, form, etc.
8.1 Introduction
Digital signature scheme (Simmons, 1984; NIST, 2000， 2006c; Lyons-Burke,
2000; Atreya, Hammond, Paine, Starrett & Wu， 2002) was firstly conjectured by
Whitfield Diffie and Martin Hellman (1976). Then Ronald Lorin Rivest, Adi Shamir
and Leonard Max Adleman (1978， 1983) realized the first digital signature scheme
using RSA algorithm, which is a type of integer factorization cryptography (IFC).
Subsequently, this classical digital signature scheme has a lot of variants carrying out
certain specific functions.
These variants are blind signature (Chaum, 1982， 1988)， multisignature
(Itakura & Nakamura, 1983; Boyd, 1989; Ham & Kresler, 1989)， group-oriented
signature (Desmedt, 1987)， undeniable signature (Chaum & van Antwerpen, 1989),
threshold signature (Desmedt & Frankel, 1989)， fail- stop signature (Pfitzmann &
Waidner, 1990; Pfitzmann, 1996)， group signature (Chaum & van Heyst, 1991),
proxy signature (Mambo, Usuda & Okamoto, 1996), signcryption (Zheng, 1997)， and
forward- secure signature (Bellare & Miner, 1999).
147
Cryptographic hash function is used in digital signature scheme to process a
specific digital message with variable length, and produce a unique fixed- length bit
stream called hash (Dang, 2007a, 2007b; Wikipedia Contributors, 2008o， 2008w) to
represent the message. Since hash is like a summary of message, it is also known as
message digest and digital fingerprint. Hash function is a one-way function with the
properties of preimage resistance and collision resistance. MD5 (Rivest, 1992;
Wikipedia Contributors, 2008r)， SHA-1 (Eastlake & Jones, 2001; NIST, 2002b;
Eastlake & Hansen, 2006; Wikipedia Contributors, 2008w) and RIPEMD-160 are the
most popular hash functions as of 2005 (Wikipedia Contributors, 2008o). They are
built based on the architecture of MD4. MD5 has a digest size of 128 bits and is no
longer sufficient for the minimum requirement of 160 bits by 2010. SHA-1 has been
compromised (Wang, Yin & Yu, 2005) and this hastens NIST to propose the
transition to SHA-2. SHA-2 consists of four variants: SHA-224, SHA-256, SHA-384
and SHA-512 (NIST, 2002b; Lilly, 2004; Eastlake & Hansen, 2006).
One of the many applications of secret is to assign a particular message with
particular object like meaning, function, or recipient. For instance (Cox, Miller,
Bloom & Fridrich, 2006)， to prevent and trace the public disclosure of government
documents by the press, Margaret Thatcher, who was British former Prime Minister
in the 1980s, inserted certain unique number of white spaces (aka blanks) as secret in
documents distributed to different cabinet ministers, so as to identify the document
recipients who have disclosed the documents to the press. This is a type of covert
text watermarking with recipient-designated message. Recipients of cabinet ministers
here are designated objects for the message of distributed government documents.
Likewise, the secret of blanks can be used to represent other objects like
specific meaning and function. Anonymity and non-repudiation are two of its not yet
well-established requirements. Comparing with watermarking, digital signature has
stronger security strength in terms of randomness, integrity, and robustness.
Nevertheless, so far there is no object-designated message using digital
signature scheme. Hence, there exists a need to create object-designated signature
scheme with optional properties of anonymity and non-repudiation called "multihash
signature". There are a few applications of hash function. Here, multiple hashes are
148
processed to be generated from a single message. This is possible by using different
rounds of hash iteration. These unique hashes are then signed by a private key to
create multiple digital signatures that can be verified using the same public key. This
new variant of classical digital signature scheme has a few new functions in term of
better message management and control.
8.2 Classical Digital Signature Scheme
In a classical digital signature scheme, a message is hashed only once. Then,
a sender signs or encrypts this generated hash using his private key to produce a
digital signature. The message and digital signature are sent to one or more receivers.
To verify a digital signature, a receiver hashes the message to produce hash Hi and
decrypts the digital signature using the sender's public key to get H 2 . If Hi is the
same with H2, then the digital signature is verified; or else, it is rejected. The
pseudocode of this classical scheme is shown in Figure 8.1.
0.0 Initialization
0. 1 Sender has private key K Pte and public key K Pub .
0.2 There may be one or more receivers.
1.0 Signing a message
1.1 Sender hashes a message using hash function to get a hash H.
1.2 Sender signs or encrypts the hash H using K Pte to get a digital signature S.
1.3 Sender sends the message and digital signature S to one or more receivers.
2.0 Verifying a message
2.1 A receiver hashes the received message using hash function to get a hash Hj.
2.2 Receiver decrypts the digital signature S using K Pub to get a hash H 2 .
2.3 Compare Hi and H 2
2.3.1 If Hi is the same with H 2 , the digital signature S is verified;
2.3.2 else if Hi is not the same with H 2 ， the digital signature S is rejected.
Figure 8.1 Pseudocode of the classical digital signature scheme
8.3 Applications of Hashing
Hash function has a few applications. It is used to hash a password or key to
form a hashed password to be stored in a computer for future authentication purpose.
149
In another password authentication scheme, hashing is used in password-
authenticated key exchange (PAKE) like SPEKE and SRP-6. Hashing is also used to
process a digital file to get a fixed- length bit stream uniquely representing this digital
file. It has the cryptographic properties of integrity and confidentiality in this case.
Any change to the message can change the hash value. Disclosing a message digest
does not reveals the contents of a message, but proves that an owner has this message
at an earlier time.
The message digest encrypted by a sender's private key produces a digital
signature that can be verified using the sender's public key. This is the classical
digital signature scheme. Further application of this classical scheme gives birth to
digital timestamping scheme (Haber & Stornetta, 1990， 1991). On the other hand,
hashing can be used for pseudorandom number generation as well.
There are some special applications of hashing. Key strengthening, wmch is
also called key stretching, is one of them (Kelsey, Schneier, Hall & Wagner, 1997).
A message like password or key is hashed many rounds. It is used to make a weak
key stronger by setting the key processing time at a maximum of one second. Some
variants of this key strengthening allow a master key to be concatenated with website
name to produce multiple site keys (aka slave keys) for different offline and online
accounts. There may be one or two levels of key strengthening. Some of the
examples are LPWA (Lucent Personal Web Assistant), HP Site Password, CPG
(Compass Password Generator), Password Multiplier, SPP (Single Password
Protocol), PwdHash, Passpet, and Multihash Key (Lee & Ewe, 2007a).
For our proposed digital signature scheme, the similar concept used in
multihash key is applied again. Instead of generating multiple hashes from a master
key, this time multiple hashes are created from a single message. This is the main
reason why this signature variant is called multihash signature.
8.4 Proposed Method: Single Message with Multiple Digital Signatures from
Single Asymmetric Key Pair
In multihash signature scheme, majority of the steps in classical digital
signature scheme are maintained. The main difference is that a message is hashed
150
using different number of hash iteration to produce multiple hashes. These hashes or
message digests are unique among themselves. When these message digests are
encrypted or signed using a sender's private key, different digital signatures are
created, but all the produced digital signatures can be verified using the sender's
public key. Figure 8.2 illustrates the pseudocode of this new signature variant.
0.0 Initialization
0. 1 Sender has private key K Pte and public key K Pub .
0.2 There may be one or more receivers with a maximum.
0.3 Sender keeps a table matching the numbers of hash iteration N to every
receiver.
1.0 Signing a message
1.1 Sender hashes a message using hash function for N rounds to get a hash H N .
1.2 Sender signs or encrypts the hash H N using K Pte to get a digital signature S N .
1.3 Sender sends the message and digital signature S N to receiver R N .
2.0 Verifying a message
2.1 Receiver R N hashes the received message for N rounds to get a hash H N1 .
2.2 Receiver decrypts the digital signature S N using K Pub to get a hash H N2 .
2.3 Compare H N1 and H N2
2.3.1 If H N i is the same with H N2 , the digital signature S N is verified;
2.3.2 else if H N1 is not the same with H N2 ， the digital signature S N is rejected.
Figure 8.2 Pseudocode of the proposed multihash signature scheme
The sender sends the message and one of the digital signatures to every
receiver. Each receiver shall receive the same message but different digital signature.
Every digital signature has a different number of hash iterations. This hash iteration
number is concatenated to the digital signature file to ease the receiver in producing a
message digest for verification purpose. Every receiver has a different hash iteration
number, and they are indexed using this number by the sender in order to identify
them. The sender keeps a table that matches the hash iteration numbers to the
identities of receivers.
From the experiments on number of hash iterations with maximum one-
second message processing time, the following results for two sets of computer
systems are obtained. For the first computer system, it is a desktop PC with Pentium
151
II 266MHz 192MB RAM running on Windows XP Professional Edition. The upper
bound for one- second hash iteration is 8200. It means 8192 (= 2 13 ) receivers can be
supported using a 13-bit string to be concatenated to the digital signature.
For the second computer system, it is a laptop PC with Centrino Duo
1.66GHz 1.5GB RAM running on Windows XP Home Edition. Its upper bound for
one- second hash iteration is 93700. In other words, 65536 (= 2 16 ) receivers can be
supported using a 16-bit string. Yet for the third computer system, it is a desktop PC
with Pentium D 2.80GHz 512MB RAM running on Windows XP Professional
Edition SP3. Its upper bound for one- second hash iteration is 122,500. In other words,
65536 (= 2 16 ) receivers can be supported using a 16-bit string.
Better computer systems with better performance can support larger number
of receivers. However, to have one- second message processing for all the computer
systems, the oldest computer system that is still in the market has to be the main
parameter in setting the maximum number of receivers.
In short, multihash signature method and system can provide object-
designated signature message with specific meaning, function, or recipient as
illustrated in Figure 8.3. A message is hashed iteratively for variable rounds by a
signor, and later signed using signor's asymmetric private key to generate a new type
of digital signature. This new digital signature only differs from the conventional
digital signature in the aspect that it carries the information of hash iteration number
as well. In other words, a message can have multiple digital signatures from an
asymmetric key pair, and each hash iteration number can be designated for any
object, action, feature, function, meaning, recipient, etc., as a representation. Here,
the signor keeps a table matching the hash iteration number and its represented object.
8.5 Comparisons for Advantages
Table 8.1 shows the main differences between the classical and multihash
signature schemes. The proposed variant has the same security strength as classical
scheme. The key difference is a message is hashed for one round in classical digital
signature scheme; whereas a message is hashed for variable round in multihash-
based digital signature scheme. This allows our signature variant to have variable
152
number of digital signatures per message, where classical scheme has one digital
signature per message.
2200
V
Settings of multihash signature to provide object-designated signature message:
(1) Signor S has an asymmetric key pair of private key K pte and public key K pub
(2) There may be one or more designated objects with a maximum like signee (or
signature receiver), action, feature, function, meaning, etc.
(3) Signor keeps a table matching the numbers of hash iteration N to each
designated object N
V
Signor S signing a message M:
(1) Signor S hashes a message M using a hash function for N rounds to get a hash
value H N
H N Hash (M ， N)
(2) Signor S signs or encrypts the H N using K pte to get a digital signature S N
S N Sign (H N , K p te)
(3) Signor S sends the message M and signature S N to signee R N
[M ， S N ] sent to R N
V
Signee R N or other parties verifying a signature message:
(1) Signee R N receives message Mi and digital signature S N i from the signor
[M! , S N1 ] ᅳ [M , S N ] ᄂ
(2) Signee R N hashes the Mi for N rounds to get a hash value H Ni
(3) Signee R N decrypts the S N i using K pub to get a hash value H N2
(4) Signee R N compares H N1 and H N2 :
if H N1 = H N2 ， digital signature S N i is verified to be signature of Mi ；
else if H N i 4- H N2 ， digital signature S N1 is rejected . end
(5) Signee R N signs S N i using one's private key K pteR to create acknowledgment
message M ack for recipient non-repudiation, and sends M ack to the signor S
V
Signor verifying an ooject-designated signature message :
(1) Signor S receives message Mu and digital signature S N u from somewhere
(2) Signor S hashes the Mu for N rounds to get a hash value H Nm
(3) Signor S decrypts the S NU using K pub to get a hash value H NU2
(4) Signor S compares H NU i and H NU2 :
if H NU i = H NU 2 ， digital signature S N u is verified to be signature of Mu
else if H NU i ^ H NU2 ， digital signature S N u is rejected . end
(5) if S N u is verified, then received Mu and S N u are from signee R N
Figure 8.3 Multihash signature method and system to provide object-designated
signature message
153
2202
Table 8.1 Comparisons of classical and multihash-based digital signature schemes
Difference
Classical
Multihash Signature
Number of hash iterations
One
Variable
Number of digital signatures per message
One
Variable
To trace a rile downloaded rrom dirrerent mirror
sites
No
Yes
To referee an advertiser broadcasting the news or
a sponsor
No
Yes
To monitor the leaking source that publicly
discloses a classified digital file
No
Yes
Security compared with classical digital
signature scheme
Same
These two main differences create three further differences that act as new
functions to digital signature scheme in term of message management and control.
Firstly, this proposed signature variant can trace a file downloaded from different
mirror sites. For example, current executable s in some websites are EXE files listed
together with .SIG files. .SIG file is a digital signature file format that allows users
who have downloaded the executable to verify the integrity of this executable.
Normally, there is more than one link for download, and these links are sharing the
same .EXE and .SIG. It would be better if we have a single .EXE with different .SIG
riles for different mirror sites. An administrator can create a table matching the hash
indices with respective mirror sites in the process of creating the multiple digital
signatures. This allows the administrator to trace the mirror site of an executaole.
Secondly, this signature variant can referee an advertiser broadcasting the
news of a sponsor. Let say someone acting as a sponsor wants to advertise news like
job recruitment, where an advertiser acting as a broadcaster of this news can earn an
allowance if an audience is successfully acquired to certain degree like interviewed
or recruited. If this sponsor does not want the information of the advertiser to be
displayed in the message file, then the sponsor can opt to have multiple digital
154
signatures for a single advertisement message. The number of hash iterations in the
digital signature file is matched to the identity of an advertiser. This matching table is
kept by the sponsor. In short, a sponsor can trace the performance of each advertiser,
who acts as a referee, by using an exactly unique message without having any
reference number in the message, where any difference in the message file can cause
ambiguity to the receivers like future employees and customers.
Thirdly, it can monitor the leaking source that publicly discloses a classified
digital file such as will, contract, form, etc. For instance, in a will, the involved
parties are will holder, lawyer(s), and witnesses. Every lawyer and every witness are
issued a different set of digital signatures with different hash iteration number to
identify them. The will holder keeps this matching table. Whenever there is a public
leakage of the will before the death of the will holder or any allowed period, the will
holder or any other party can identify who is the leaking source.
In short, advantages of multihash signature are designated recipient function
to alternate with watermarking, object-designated meaning, referral function,
anonymity support, avoidance of name clashing and renaming problems, stronger
collision resistance than method using the hashing of the concatenation of message
digest and object name like Equation (8.1)， as well as recipient non-repudiation. The
example of object-designated meaning is the cheque validity status including status
like valid, invalid, paid, void, on hold, late processing, rejected, withdrawn, cancelled,
etc. The examples of referral functions are to trace a file downloaded from different
websites, to referee an advertiser broadcasting the news of a sponsor, and to monitor
the leaking source that has publicly disclosed a classified digital file.
Hash Value = Hash(Hash(Message) II Object Name) (8.1)
Here, multihash signature is used in some other inventions of this doctoral
research project. One of them is called triple-watermark digital cheque and another is
triple-watermark software licensing schemes in Chapter 10， together with MePKC,
steganography, and watermarking. The security of multihash signature has the same
strength with the conventional digital signature scheme. For higher security to trace
the identity of an Internet user signing a message and one's Internet geographical
155
region, a message is suggested to be hashed and concatenated with MAC address
and/or IP address, and then undergoes an optional conventional digital signature or
multihash signature as in Equation (8.2) where S = Signature.
S = Multihash Signature (Hash (Message) II MAC Address II IP Address) (8.2)
8.6 Conclusion
In a nutshell, introducing variable round of hash iterations on a message can
create new functions to digital signature scheme. Using multiple digital signatures
from single message, which are sent to different receivers, can trace the file
downloaded from some mirror sites, referee an advertiser broadcasting the
advertisement of a sponsor, and monitor the leakage of any classified digital file. The
digital signatures in this proposed signature variant called multihash signature are
unique among themselves. However, they can be verified using single public key
even though they are generated from single message.
156
CHAPTER 9 APPLICATIONS OF BIG SECRET & MePKC
(PART 1)
9.1 Applications of Created Big Memorizable Secret(s)
For useful applications of the created big memorizable secret(s) and MePKC
(Memorizable Public-Key Cryptography), they include:
(i) methods and systems to realize memorizable symmetric key the secret till
resistance to quantum computer attack;
{11) methods and systems to realize memorizable public-key cryptography
(MePKC);
^m) methods and systems to improve security strength of other cryptographic,
information-hiding, and non-cryptographic applications of secret beyond
128 bits;
(iv) method and system to harden the identification of embedded data in
steganography although stego-data has been detected;
(v) method and system to transfer fund electronically over a remote network
using MePKC;
^vi) method and system to license software electronically over a remote
network using MePKC;
{vii) methods and systems to authenticate human-computer and human-
human communications at a local station or over a remote network using
MePKC;
^vin) method and system to use digital certificate with more than one
asymmetric key pair for different protection periods and password
throttling;
(ix) method and system to use three-tier MePKC digital certificates for ladder
authentication;
157
(x) method and system to store, manage, and download voice and video calls
of mobile phone and wired phone at online distributed servers;
(xi) method and system of multipartite electronic commerce transactions; as
well as
(xii) Method and system to boost up the trust level of MePKC digital
certificate by using more than one certification authority (CA) and/or
introducer of trust of web.
Applications (i-iii) are presented in this Chapter 9. Applications (iv-vi) are
presented in Chapter 10. Applications (,vii-ix) are explained in Chapter 11. For
applications (x-xii), they are in Chapter 12.
To apply big memorizable secret(s) to the novel methods and systems using
MePKC from (iv) to (xii), two more independent inventions are applied here to
enhance the features of MePKC. These two inventions are multihash key and
multihash signature (aka object-designated signature). Multihash key includes some
methods and systems to generate multiple slave keys from a single master key as in
Chapter 7. Meanwhile, multihash signature includes a method and system to generate
object-designated signature message with specific feature, meaning, function, or
recipient as in Chapter 8.
9.2 Memorizable Symmetric Key to Resist Quantum Computer Attack
Due to the successful cracking of 56-bit DES (Data Encryption Standard) in
the 1990s, stronger symmetric ciphers with larger symmetric key sizes like 80-bit
2TDES, 112-bit 3TDES, as well as 128-, 192-, and 256-bit AES (developed from
Rijndael cipher) are introduced to replace the DES.
Blaze, Diffie, Rivest, Schneier, Shimomura, Thompson and Wiener (1996)
discussed the minimal key lengths for symmetric ciphers. The NIST (National
Institute of Standards and Technology), USA, proposes different protection periods
for security through years 2010， 2030， and beyond 2030， for 80， 112, and 128 bits,
respectively (E. Barker, W. Barker, Burr, Polk, & Smid, 2007a, 2007b). ECRYPT of
158
European Union (EU) proposes in its technical reports that 80-， 96-， 112-， 128-， and
256-bit security have protection periods of 4 years through year 2010， 10， 20， 30
years, and foreseeable future to be against quantum computer attack, respectively
(Gehrmann & Naslund, 2005, 2006， 2007). Nevertheless, conventional methods and
systems normally can only realize a key size of 128 bits or less.
Hence, the first application 9.1(i) of the present invention in applying the
created big memorizable secret is to realize higher security levels of symmetric
ciphers like AES-192 and AES-256. By using the methods and systems as in Figure
3.1 and Table 3.3, it can be observed that the current highest security level of
symmetric cipher at 2^o bits can be practically realized and achieved using big
memorizable 256-bit secret. For another onef explanation, please visit Section 4.3.9.
9.3 Memorizable Public-Key Cryptography (MePKC)
9.3.1 Related Works: Storages of Private Key
For the current asymmetric key crypto system, a private key is normally
encrypted using another symmetric key. The encrypted private key is storea in a
local computing system or token; whereas the symmetric key is stored in the human
brain. The present possible attacks for this method are guessing attack, dictionary
attack, and pre-computation attack.
Another method is to split the private key into two or more portions (Ganesan,
1996b; Bishop, 2003， pp. 264-265). There are other literary works about split private
key cryptography over here (Ganesan, 1996a, 1998a, 1998b, 1998c, 1999; Ganesan,
Sandhu, Cottrell & Austin, 2006a, 2006b, 2006c; Ganesan, Sandhu, Cottrell,
Schoppert & Bellare, 2006; Ganesan & Yacobi, 1996; Sandhu, deSa & Ganesan,
2003， 2005a, 2005b, 2005c, 2006a, 2006b, 2006c, 2006d， 2006e， 2006f; Sandhu,
Schoppert, Ganesan, Bellare & deSa, 2006a, 2006b, 2006c, 2006d， 2006e， 2006f，
2007a, 2007b, 2007c, 2007d; Sandhu, Ganesan, Cottrell, Renshaw, Schoppert &
Austin, 2007; ). The first portion of the private key can be derived from a normal
human- memorizable symmetric key. The other portions of the private key are stored
as encrypted partial private key alike the normal encrypted private key. This method
resists the pre-computation attack.
159
A third method is to store the encrypted private key in a server connected to a
computer communication network (Baltzley, 2000， 2001a, 2001b). A user has the
roaming capability where the encrypted private key can be downloaded from the
server for decryption at anywhere. Proxy servers are needed for this method to avoid
single point of failure. Its possible attacks are the same as encrypted private key
stored in the local computing system.
9.3.2 The Proposed MePKC Appliacation 9.1(H)
The second application 9.1(ii) of the present invention in applying the created
big memorizable secret is to improve from the token-based public-key cryptography
(PKC) to the realization of secret-based PKC using fully memorizable private key,
which is named as MePKC (Memorizable Public-Key Cryptography) or MoPKC
(Mobile Public-Key Cryptography) here. The main advantages of MePKC are full
secret memorizability and mobility convenience. Yet another quite important
advantage is that secret-based MePKC can resist some side-channel attacks
vulnerable to token-based PKC, such as those attacks over the fully or partially
encrypted private key. For illustration of MePKC, please refer to Figure 9.1. For
another brief explanation, please visit Section 4.3.9.
The current lowest key size requirement of asymmetric private key is 160 bits
operating in FFC and ECC. From Table 3.3 listing all the proposed novel methods
and systems to create big memorizable secret, a 160-bit secret for 160-bit fully
memorizable private key can be supported by self-created signature-like Han
character for CLPW and CLPP, 2D key, multilingual key, and multi-tier geo-image
key. This group of big memorizable secret creation method and system can easily
support memorizable private key up to 256 bits at the symmetric bits of security
strength of 128 bits and for a protection period of 30 years.
For higher security levels up to 512-bit secret used by 512-bit MePKC, multi-
factor multimedia key using software token has to be adopted to halve the key size
requirement towards a practical realization. Here, the mobility convenience is
somehow sacrificed.
160
1300
V
Optionally activate the anti-keylogging software.
1301
Y
Open the MePKC application software operating on at least 160-
bit ECC (Elliptic Curve Cryptography) for its input interface.
/X 끼
1302
User creates an n-bit secret S like 256 bits using one or more methods as follows:
(1) Self-created signature-like Han character for CLPW and later CLPP
(2) ASCII-based 2D key
(3) Unicode-based 2D key
(4) Multilingual key
(5) Multi-tier geo-image key
(6) Conventional secret creation methods and other future methods
1303
User creates an asymmetric key pair as follows:
(1) Let Kpte = private key, K pub = public key
(2) Kpte Box 404 (S)， optional secret processing of memorizable secret S
(3) K pub Public Key Generation (K pte )
(4) Store the K pub and clear K pte in the computer memory
(5) Create pubUc key certificate (aka digital certificate) from K pub using certificate
authority or introducer of web of trust
(6) Optionally publish and/or send the public key certificate to other PKC users
1304
Apply the asymmetric key pair and public key certificate for
various MePKC applications like encryption, signature, etc.
1305
V
Clear the memory storing all forms of secrets. Then, close all the
application software.
1306
Figure 9.1 Operation of MePKC method and system
To generate this software token, firstly a Dig multimedia data file like random
or non-random bit stream, text, image, audio, animation, or video, is hashed by a 211-
bit hash function to produce 2n-bit hash value. The 2n-bit hash value is encrypted by
using an n-bit symmetric key and n-bit AES to further produce a software token.
161
Then, the multimedia data file is destroyed or hide at a safe location like safety box,
and the software token is either stored in a local storage device like USB flash drive
or in a remote server accessible through roaming network.
A user remembers only the n-bit secret of symmetric key. Whenever 2n-bit
MePKC is needed for various applications, the software token is acquired and
decrypted using the n-bit memorizable secret of symmetric key to obtain the 2n-bit
hash value. This n-bit secret and 2n-bit hash value are then used to derive the 2n-bit
MePKC private key.
The MePKC can be used for major PKC cryptographic applications like
encryption and digital signature schemes. Other minor applied cryptographic
schemes are key exchange, authentication, blind signature, multisignature, group-
oriented signature, undeniable signature, threshold signature, fail- stop signature,
group signature, proxy signature, signcryption, forward- secure signature, designated-
verifier signature, public-key certificate (digital certificate), digital timestamping,
copy protection, software licensing, digital cheque (aka electronic cheque), electronic
cash, electronic voting, BAP (Byzantine Agreement Protocol), electronic commerce,
MAC (Message Authentication Code), key escrow, online verification of credit card,
multihash signature, etc.
The blind signature scheme includes its further applications for electronic
cash (aka e-cash, electronic money, e-money, electronic currency, e-currency, digital
cash, digital money, digital currency, or scrip), and electronic voting (aka e-voting,
electronic election, e-election, electronic poll, e-poll, digital voting, digital election,
or digital poll).
Advancement of computing technologies requests for longer key sizes for a
fixed protection period. To freeze this unwanted request, key strengthening (aka key
stretching) through many rounds of hash iteration, together with hash truncation and
a hash function with longer hash value like 768， 1024 bits or more, can be used.
MePKC is extended to a novel claimed invention here called multihash
signature scheme, and novel innovations of some cryptographic schemes like digital
cheque, software licensing, human-computer and human- human authentication via a
162
computer communications network, as well as MePKC digital certificate with
multiple public keys for password throttling and ladder authentication.
These MePKC applications are best to be implemented using the ECC
(Silverman, 1986; Blake, Seroussi & Smart, 1999， 2005; Hankerson, Menezes &
Vanstone, 2004， 2005; Zhu & Zhang, 2006). This is because ECC needs a minimum
private key size of 160 bits and it has been long time tested for its security strength.
Alternatively, depending on further research and evaluation, shorter private key size
at equivalent or better bits of security strength can be achieved by using hyperelliptic
curve cryptography (HECC) (Pelzl, Wollinger & Paar, 2004; Cohen & Frey, 2006;
Wang & Pei, 2006) and possibly other crypto systems like torus-based cryptography
(TBC) (Rubin & Silverberg, 2003).
For HECU, the genera 2 and 3 have so far been tested to have shorter key size
requirement than ECC by twice and thrice. Between them, genus-2 HECC has a
higher security without the demand to have a correction factor for its key size. In
other words, the correction factor of HECC of genus 2 is 1. As information, genus-3
and genus-4 HECC have a correction factor of 1.05 and 1.286 times of its field,
respectively, for the key size to get a larger group order at equivalent bits of security
strength. For more information, please refer to an article entitled "High Performance
Arithmetic for Special Hyperelliptic Curve Crypto systems of Genus Two" by Jan
Pelzl, Thomas Wollinger, and ChristofPaar (2004).
9.4 Other Cryptographic, Information-Hiding, and Non-Cryptographic
Applications of Secret beyond 128 bits
The third application 9.1(m) of the present invention in applying the created
big memorizable secret is various other cryptographic, information- hiding, and non-
cryopto graphic applications needing a big memorizable secret(s).
The other cryptographic applications include various PAKE (Password-
Authenitcated Key Exchange) like SPEKE (Simple Password Exponential Key
Exchange) (Jablon, 2006) and SRP-6 (Secure Remote Password Protocol version 6)
(Wu， 2003).
163
Meanwhile, information- hiding applications (Petitcolas, Anderson & Kuhn,
1999; Moulin & O' Sullivan, 2003) include stego-key in steganography (Simmons,
1984, 1998; Anderson & Petitcolas, 1998; Cachin, 1998; Mittelholzer, 1999; Fridrich
& Goljan, 2004; Fridrich, Goljan & Soukal, 2004; Lu， 2005), secret key in
symmetric watermarking, and private key in asymmetric watermarking (Swanson,
Kobayashi & Tewfik, 1998; Low & Maxemchuk, 1998; Hartung & Kutter, 1999;
Mittelholzer, 1999; Mohanty, 1999; Wolfgang, Podilchuk & Delp， 1999; Eggers, Su
& Girod, 2000; Collberg & Thomborson, 2002; Hachez & Quisquater, 2002; Arnold,
Schmucker & Wolthusen 2003; Furon & Duhamel, 2003; Barni & Bartolini, 2004;
Furon, 2005; Cayre Fontaine & Furon, 2005a, 2005b, 2005c; Lu， 2005; Cox, Doerr
& Furon, 2006; Furht & Kirovski, 2006a, 2006b).
Lastly, non-cryptographic applications include seed for PRNG (Pseudo-
Random Number Generator) and CSPRBG (Cryptographically Secure Pseudo-
Random Bit Generator) (Eastlake, Crocker & Schiller, 1994; Rukhin, Soto,
Nechvatal, Smid, Barker, Leigh, Levenson, Vangel, Banks, Heckert, Dray & Vo，
2001; Le Quere, 2004; Keller, 2005; Barker & Kelsey, 2007; Campbell & Easter,
2007b) like the Blum-Blum-Shub (BBS) CSPRNG (Mollin, 2007a, p. 508, 2007b).
164
CHAPTER 10 APPLICATIONS OF BIG SECRET & MePKC
(PART 2)
10.1 Identification Hardening of Embedded Data in Steganography
10.1.1 Related Works
Steganography is a branch of information hiding. Secret message acts as
embedded data into a cover data under the control of a stego-key to form a stego-data.
Stego-data in its forms of storage and transmission through an insecure channel shall
be like a normal data without triggering the suspicion of a person sensing the stego-
data. To retrieve the secret message, the stego-data is processed using the stego-key
to get back the embedded data.
In the current prior art, reliable detection of stego-image can be done
successfully as shown by Fridrich and Goljan (2004). Yet the stego-key searching
can also be done within promising time for a short stego-key.
This is reported by Fridrich, Goljan, and Soukal (2004) in "Searching for the
Stego-Key" that as long as the embedded message is not occupying 100% of image
capacity, then stego-key searching is independent of encryption key and takes about
12 hours to crack a 30-bit stego-key. Hence, there exists a need to have a big and yet
memorizable stego-key, and to somehow fully occupy the data capacity for higher
complexity to resist the cracking of steganographic system.
10.1.2 The Proposed MePKC Application 9.1(iv)
The fourth application 9.1(iv) of the present invention in applying the created
big memorizable secret is to boost up the key size of stego-key to be more than 128
bits. Based on extrapolation of an article by Fridrich, Goljan, and Soukal (2004)， for
an 80-bit stego-key, it has a protection period of about 5 years or usable by year 2010
alike the 80-bit symmetric key. It is the contribution of the present embodiment to
harden the identification of embedded data in steganography even after the stego-
data has been detected as in Figures 10.1-10.2.
165
2300
V
Required components to harden the identification of embedded data in steganography:
(1) Steganosystem where sender and receiver of a stego-data shared a stego-key
(2) Symmetric key cryptosystem like AES-256
(3) Asymmetric key cryptosystem like 512-bit MePKC operating on ECC
(4) CSPRBG (Cryptographically Secure Pseudo-Random Bit Generator)
(5) Lossless multimedia data compression like BMP, PNG, and TIFF for image
V
Preparing ciphertext of embedded secret data M and symmetric key K SY ：
(1) Generate an n-bit random number as a symmetric key K SY ， where n = 256
(2) Encrypt the embedded data M using K SY under AES-256 to produce ciphertext Cj
(3) Encrypt the K SY using recipient' s public key K pub to produce NL-bit ciphertext Ck
where N L = 512
V
Creating a stego-data by embedding secret message into cover-data:
(1) Seed an N S T-bit stego-key K ST into a CSPRBG to produce sequential units of N R -bit
bitstream B， where Nst = 256 and Nr = 32
(2) Assume the cover data is a PNG image with dimensions (x * y) and bit depth per
channel at B P bits for channels RGBA, where x = y= 1024, B P = 8， N P = number of
bits/pixel = 32, then S S i ze = maximum supported size of embedded data in a cover data
= ぶ * y * Bp = 1024 * 1024 * 8 > total size of C M and C K
(3) Every pixel of the image is indexed by an address location starting from the top
leftmost pixel, moving to the rightmost pixel, and then continuing with the leftmost
pixel of the second line, and so on, until the rightmost pixel in the last bottom line
(4) For every sequential unit of NR-bit bitstream B， calculate L P = (B mod (x * y)) to get
the selected pixel location in the cover image, where L P = B mod 2 20 , and first, second,
third, and so on of the B are labeled as B ， B } , B 2 , -. -, B N
(5) For every B N ， record it into an index table, and if a B N has occurred previously, mark
and use the subsequent (B N + 1) as the selected pixel location
(6) Chunk the Ck and Cm into Bp-bit block, and store the chunks of Ck first, followed by
chunks of Cm, one by one, into the Bp-bit alpha channels addressed by the NR-bit
bitstream B to produce a partially completed stego-data
V
Creating a stego-data with data capacity fully occupied where for example data is image:
(1) Seed another CSPRBG with the present clock time to produce sequential garbage
units of Bp-bit bitstream G to harden the identification of embedded data
(2) Store G addressed by additional NR-bit bitstream B into the remaining alpha channels
of remaining pixel locations until the index table has all the pixel locations marked
Figure 10.1 Data embedding process into a cover data for method and system to
harden the identification of an embedded data in steganography although stego-data
has been detected
166
〜蕭
2302
2304
2400
V
Required components to extract the embedded data from the hardened stego-data:
(1) Steganosystem where sender and receiver of a stego-data shared a stego-key
(2) Symmetric key cryptosystem like AES-256
(3) Asymmetric key cryptosystem like 512-bit MePKC operating on ECC
(4) CSPRBG (Cryptographically Secure Pseudo-Random Bit Generator)
(5) Lossless multimedia data compression like BMP, PNG, and TIFF for image
V
Calculating the embedded sequences of symmetric key Ks Y and embedded secret
data M:
(1) Use NsT-bit stego-key Ks T to generate sequential units of N R -bit bitstream B
(2) Calculate L P = (B mod (x * y)) and its subsequent value if there is a clash to
get the series of selected pixel locations in the stego-image
(3) Extract the ciphertext C K
(4) Extract the ciphertext C M
2402
、
Decrypting the ciphertexts of symmetric key C K and embedded secret data C M :
(1) Decrypt the ciphertext C K using the recipient's private key K pte to get
symmetric key Ks Y
(2) Decrypt the ciphertext C M using the K SY to retrieve the embedded data M
、
，
Clear the memory storing all forms of secrets and close all the application
software.
2403
2404
Figure 10.2 Data extracting process of embedded data from a stego-data for method
and system to harden the identification of an embedded data in steganography
although stego-data has been detected
Firstly, a stego-key is shared between the sender and receiver using some key
exchange protocol like PAKE and MePKC key exchange scheme. Then, a symmetric
key is created from a CSPRBG and use it to encrypt an embedded secret data to
produce ciphertext of embedded data Cm. The symmetric key is later encrypted by
recipient's public key to produce ciphertext of symmetric key Ck. To identify the
address locations to hide the Cm and Ck, another CSPRBG is seeded with the stego-
key and used to produce a list of addresses. Every unique address is recorded in an
167
index table. If a generated address clashes with an address in the index table, then its
subsequent address not in the index table is used.
After the Cm and Ck are hidden into the cover data, then use a third CSPRBG
to generate random garbage bit streams G and use them to fully occupy the
remaining data capacity. Consequently, from the full occupation of data capacity, the
complexity to search for a stego-key will be higher when even encryption key
searching is needed for cracking. To paralyze the stego-data detection, a sender can
often broadcast dummy stego-data with noises as the embedded data.
10.2 Electronic Fund Transfer Using MePKC
10.2.1 Related Works
Among the various applications of digital signature scheme, electronic
cheque (aka digital cheque) is a special and important type of messages. Electronic
cheque (Doggett, Jaffe & Anderson, 1997; Anderson, Jaffe, Hibbert, Virkki, Kravitz,
Chang & Palmer, 2000， 2001) introduced another form of electronic fund transfer
using conventional digital signature scheme. The popularity of these method and
system are low due to the drawbacks of PKC (Public-Key Cryptography), i.e. low
mobility of partially or fully encrypted private key, and management difficulty of
certificate revocation list. Furthermore, the digital signature of Doggett' s method
carries only the information of electronic fund transfer from a payer to a payee via
one or more banks.
In fact, a physical cheque has various processing states for accounting records
like blank cheque, signed for payment, paid cheque, returned cheque by payee,
withdrawn payment by payer, withdrawn payment by payer's bank, bounced cheque,
advanced cheque, outdated cheque, fake cheque, etc.
Yet for electronic cheque, that transfer fund between accounts electronically
and speedily throughout the world in the computer network, it shall have more
optional security protection beyond the digital signature because money is a sensitive
and critical object to be tracked for the convenient investigation of (organized)
criminal activities (Lampe, No date; Glick, 1995; Livingston, 1996; "UNODC and
168
2502d {
Organized Crime," No date; Layman & Potter, 1997; Maxim & Whitehead, 1998;
Chen, 2004; Ruan & Wang, 2005; Siegel, 2005; Wang, 2007; "Identity-Related
Crime," 2007; He, 2007; United Nations Office on Drugs and Crime (UNODC),
2004, 2008; Wikipedia Contributors, 2008s, 2008af) and civil cases.
2500
People Bank Cheque No.: ABC123 Payer's Name: Kok-Wah Lee
Kampsu-, Perak, Malaysia I C /Passport: MY1234S67890
Email: che que @p e opleta ank . c om.my Email: kokwah@xpreeli.com
2501
People Bank
Cheque No.: ABC123 Payer's Name: Kok-Wah Lee
Kampar, Perak, Malaysia
Email: che que @p e opleb ank . c om.my
Payee's Name: Xpree Jinhua Li
IC/Passport: MY1 234567891
Email: xpreee@gmail.com
Pay Ringgit One Thousand Only.
IC/Passport: MY1234S67890
Email: kokwah@xpreeli.com
Date: 24 June 2008
Amount: MYRS1 000-00
Figure 10.3b Written cheque signed by payee
2502
People Banh
Cheque No.: ABC123 Payer's Name: Kok-Wah Lee
Kampar, Perak, Malaysia
Email: che que @pe opleb arik.c om.my
Payee's Name: Xpree Jinhua Li
IC/Passport: MY1 234567891
Email: xpreee@gmail.com
Pay Ringgit One Thousand Only.
IC/Passport: MY123+S67890
Email: kokwah@xpreeli.com
Date: 24 June 2008
Amount: MYRS1 000-00
2500a
>■ 2500b
> 2500c
2501a
>■ 2501b
> 2501c
2502a
y 2502b
2502c
Figure 10.3c Processed payee's cheque by bank
Figure 10.3 Samples of digital cheque in triple-watermark digital cheque scheme
169
2600
\レ
Required components for a digital cheque method and system:
(1) Symmetric and asymmetric watermarking systems
(2) Asymmetric key cryptosystem like 512-bit MePKC operating on ECC
(3) CSPRBG (Cryptographically Secure Pseudo-Random Bit Generator)
(4) Lossless multimedia data compression like BMP, PNG, and TIFF for image
V
Key exchange for a shared symmetric WM key K WM between payer and bank:
(1) Payer creates K WM using a username, random number R, and payer's private
key Kp te i
K W m Sign ( Hash (Username II R) , K ptel )
(2) Payer sends the K WM to bank using a key exchange protocol like MePKC
V
Bank preparing blank cheque for payer:
(1) Bank writes the bank (name, branch, email, etc.), payer (name, IC/passport,
email, etc.), and cheque number in a blank PNG image file as in Figure 25a
(2) For the partial image portion 2500a, hash it and then sign the hash using
bank's private key K pte0 to produce signature S
S Sign ( Hash (Image Portion 2500a) ， K pte0 )
(3) Bank embeds S as first watermark WM to the top band of image portion
2500c in red using K WM to select pixel address locations for WM embedding
as in Figure 23， where K WM acts like the stego-key
(4) Other remaining pixel locations in the red band are filled with random bits
(5) Bank sends the prepared blank cheque CHQ 2500 to a payer
V
Payer verifying, writing, and signing a digital cheque:
(1) Payer verifies WM of CHQ using K WM and bank's public key K pub0
(2) If WM is verified, payer writes the payee (name, IC/passport, email, etc.),
cheque amounts, and date to create image portion 2501b as in Figure 25b
(3) For the partial image portions 2501a and 2501b, hash them and then sign the
hash using payer's private key K ptel to produce signature Si
Si Sign ( Hash (Image Portion 2501a II Image Portion 2501b) ， K ptel )
(4) Payer embeds Si as second watermark WMi to the middle band oi image
portion 2501c in green using K WM to select pixel address locations for WMi
embedding as in Figure 23， where K WM acts like the stego-key again
(5) Other remaining pixel locations in the green band are filled with random bits
(6) Payer sends written and signed digital cheque CHQi to payee via MePKC
Figure 10.4 Creation of blank cheque by a bank and written cheque by a payer in the
triple-watermark digital cheque method and system
170
〜肅
ᅮ 2602
ᅮ 2603
From 2604
V
Payee's cheque crediting actions in a digital cheque method and system:
(1) Payee uses MePKC encryption scheme to decrypt the received digital cheque
CHQi from payer
(2) Payee uses MePKC digital signature scheme to verify the integrity of CHQi
(3) If CHQi is verified, payee sends CHQi to payer's bank or payee's bank
(4) If payee's bank, payee's bank routes CHQi to payer's bank via bank network
V
Bank processing written cheque CHQi for payer and payee:
(1) Bank verifies WMi of CHQi using K WM and payer's public key K publ
(2) If WMi is verified, bank obtains the payer's signature Si to order a payment
(3) Bank uses multihash signature to sign the image portion 2502d using bank's
private key K pte0 for an object-designated status of processed cheque like valid,
invalid, paid, void, on hold, late processing, rejected, withdrawn, cancelled,
etc., and then to produce signature S 2
S 2 Multihash Signature ( Hash (Image Portion 2502d) ， K pte o )
(4) Bank embeds S 2 as third watermark WM 2 to the bottom band oi image portion
2502c in blue using bank's asymmetric watermarking private key K WM , pte or
published symmetric WM key K WM2 to select pixel address locations for WM 2
embedding as in Figure 23， where K WM , pte or K WM2 may also act like stego-key
(5) Other remaining pixel locations in the blue band are filled with random bits
(6) Payer's bank debits the payer's account for the cheque amount
(7) Payer's or payee's bank credits the payee's account for the cheque amount
(8) Bank sends processed digital cheque CHQ 2 to payer and payee via MePKC
V
Payer verifying the processed digital cheque CHQ 2 :
(1) Payer verifies WM 2 of CHQ 2 using bank's asymmetric watermarking public
key K WM ， pub or published K WM2 ， and bank's public key K pub0
(2) If WM 2 is verified, payer checks the bank account for the debit transaction
(3) Otherwise if WM 2 is rejected, payer reports to the bank for investigation
V
Payee verifying the processed digital cheque CHQ 2 :
(1) Payee verifies WM 2 of CHQ 2 using bank's asymmetric watermarking public
key K WM ， pub or published K WM2 ， and bank's public key K pub0
(2) If WM 2 is verified, payee checks the bank account for the credit transaction
(3) Otherwise if WM 2 is rejected, payee reports to the bank for investigation
Figure 10.5 cheque crediting process by a payee in the triple-watermark digital
cheque method and system
171
〜謂
〜篇
2702
2703
Hence, there exists a need to boost the PKC popularity, to add more
embedded information, and to increase the security strength of electronic cheque, by
applying fully memorizable private key, object-oriented signature scheme (aka
multihash signature scheme), and optional fragile watermarking scheme, respectively.
10.2.2 The Proposed MePKC Application 9.1(v)
The fifth application 9.1(v) of the present invention in applying the created
big memorizable secret is a method and system to transfer fund electronically over a
remote network using MePKC, CSPRBG, lossless data compression, as well as
information- hiding techniques like steganography and fragile watermarking, as in
Figures 10.3-10.5 and called triple-watermark digital cheque method and system.
Stronger security and prettier aesthetics are needed for digital cheque that is faster,
more efficient, and more environment-friendly than paper cheque and electronic
textual cheque using PKC merely.
There are three watermarks in the digital cheque. The first watermark marks
the information of payer's bank, payer, and cheque account signed by a payer's bank.
The second watermark marks the information of payee and cheque amount signed by
a payer. The third watermark marks the cheque status after processed by the payer's
bank like valid, invalid, paid, void, on hold, late processing, rejected, withdrawn,
cancelled, etc.
To save the image size, lossless image compression file format like PNG
(Portable Network Graphics) and TIFF (Tagged Image File Format) shall be used
besides BMP (Bitmap file format). Moreover, the digital cheque can also be in the
data type of text. Also, this method and system can be modified and applied in other
fields like software licensing.
10.3 Electronic Software Licensing Using MePKC
10.3.1 Related Works
Yet in another application of PKC (Public-Key Cryptography), software
licensing is part of software copy protection besides code obfuscation against reverse
172
engineering, watermarking against software piracy, and tamper-proofing against
tampering (Collberg & Thomborson, 2002). In the current prior art, software
licensing scheme uses fully or partially encrypted private key of PKC. Token
containing the encrypted private key is subject to loss and damage; whereas server
containing the encrypted private key is subject to virtual hacking and subsequently
guessing attack, dictionary attack, and pre-computation attack.
For computer software, its representative monetary value is its software
product ID key rather than the duplicable electronic executable and storage device
like floppy disk, CD-ROM, DVD, BD, HD DVD, etc., that stores the executable.
Hence, there exists a need for current software licensing scheme to apply the fully
memorizable private key for higher security and mobility, as well as to add more
information using multihash signature scheme, and to have extra optional security
protection to the software product ID key by using the fragile watermarking scheme.
10.3.2 The Proposed MePKC Appliacation 9.1(vi)
The sixth application 9.1(vi) of the present invention in applying the created
big memorizable secret is a method and system to license software (Manoharan &
Wu， 2007) electronically over a remote network using MePKC, CSPRBG, lossless
data compression, as well as information-hiding techniques like steganography and
fragile watermarking, as in Figures 10.6-10.8 and called triple-watermark digital
software license method and system. Ethics, self-discipline, and education are mostly
needed to fight against the software piracy (Limayem, Khalifa & Chin, 2004).
There are three watermarks in the digital software license. The first
watermark marks the information of software licensing vendor, reseller (or sales
agent), and reseller' s account signed by a vendor. The second watermark marks the
information of licensee and license selling price signed by a reseller. The third
watermark marks the software license status after processed by the vendor like
granted, upgraded, resold, void, withdrawn, evaluation, transferred, etc.
To save the image size, lossless image compression nle format like PNG
(Portable Network Graphics) and TIFF (Tagged Image File Format) shall be used
besides BMP (Bitmap file format). Moreover, the digital software license can also be
173
text data type. Also, this method and system can be modified and applied in other
fields like digital cheque.
2800
Xpreeli Enterprise
Kampar, Perak, Malaysia
Email: kokwah@xpreeli.com
License No.: ABC123 Signee's Name: Kok-Wah Lee
IC/Passport: MY1234567S90
Email: kokwoh@xpreeli.com
Figure 10.6a Blank software license issued by software vendor to reseller
2801
Xpreeli Enterprise License No.: ABC123 Signee's Name: Kok-Wah Lee
Kampar, Perak, Malaysia IC/Passport: MY1234 56 7 S90
Email: kokwah@xpreeli.com Email: kokwoh@xpreeli.com
Licensee's Name: Xpree Jinhua Li Date - 24 June 2008
IC/Passport: MY1 234567891
Email: xpreee@gmail.com AmOUlltl MYR$1 000-00
A license copy of MobileECC 1.2 software.
Figure 10.6b Written software license signed by reseller (or sales agent)
2802
Xpreeli Enterprise License No.
Kampar, Perak, Malaysia
Email: kokwah@xpreeli.com
Licensee's Name: Xpree Jinhua Li
IC/Passport: MY1 234567891
Email: xpreee@gmail.com
A license copy of MobileECC 1.2 software.
ABC123 Signee's Name: Kok-Wah Lee
IC/Passport: MY1234567S90
Email: kokwah@xpreeli.com
Date: 24 June 2008
Amount: MYRS1 000-00
Figure 10.6c Processed software license by vendor
Figure 10.6 Samples of digital software license in triple-watermark digital software
license scheme
174
2900
\レ
Required components for a digital software licensing method and system:
(1) Symmetric and asymmetric watermarking systems
(2) Asymmetric key cryptosystem like 512-bit MePKC operating on ECC
(3) CSPRBG (Cryptographically Secure Pseudo-Random Bit Generator)
(4) Lossless multimedia data compression like BMP, PNG, and TIFF for image
N レ
Key exchange for a shared symmetric WM key K WM between reseller and vendor:
(1) Reseller creates K WM using a username, random number R, and reseller's
private key K ptel
K W m Sign ( Hash (Username II R) , K ptel )
(2) Reseller sends the K WM to vendor using a key exchange protocol like MePKC
N レ
Software vendor preparing blank software license for reseller or sales agent:
(1) Vendor writes the vendor (name, email, etc.), reseller (name, IC/passport,
email, etc.), and license number in a blank PNG image file as in Figure 28a
(2) For the partial image portion 2800a, hash it and then sign the hash using
vendor's private key K pte0 to produce signature S
So ᅳ— Sign ( Hash (Image Portion 2800a) ， K pte0 )
(3) Vendor embeds S as first watermark WM to the top band oi image portion
2500c in red using K WM to select pixel address locations for WM embedding
as in Figure 23， where K WM acts uke the stego-key
(4) Other remaining pixel locations in the red band are filled with random bits
(5) Vendor sends the prepared blank software license SLC 2800 to a reseller
V
Reseller or sales agent veriiying, writing and signing a digital software license:
(1) Reseller verifies WM of SLC using K WM and vendor's public key K pub0
(2) If WM is verified, reseller writes the licensee (name, IC/passport, email,
etc.), payment, and date to create image portion 2801b as in Figure 28b
(3) For the partial image portions 2801a and 2801b, hash them and then sign the
hash using reseller's private key K ptel to produce signature Si
Si Sign ( Hash (Image Portion 2801a II Image Portion 2801b) ， K ptel )
(4) Reseller embeds Si as second watermark WMi to the middle band oi image
portion 2801c in green using K WM to select pixel address locations for WMi
embedding as in Figure 23， where K WM acts like the stego-key again
(5) Other remaining pixel locations in the green band are filled with random bits
(6) Reseller sends written and signed SLCi to licensee via MePKC
Figure 10.7 Creation of blank software license by a vendor and written software
license by a reseller in the triple-watermark digital software license method and
system
175
〜雇
2902
2904
/XT'
From 2904
V
Licensee's endorsement actions in a digital software license method and system:
(1) Licensee uses MePKC encryption scheme to decrypt the received digital
software license SLCi from reseller
(2) Licensee uses MePKC digital signature scheme to verify the integrity of SLC
(3) If SLCi is verified, licensee sends SLCi to software vendor or licensor
(4) If not software licensing vendor (SLV), other vendor routes SLCi to SLV
V
SLV vendor processing written software license SLCi for reseller and licensee:
(1) Vendor verifies WMi of SLCi using K WM and reseller's public key K publ
(2) If WMi is verified, vendor obtains reseller's signature Si for an endorsement
(3) Vendor uses multihash signature to sign the image portion 2802d using
vendor's private key K pte0 for an object-designated status of processed
software license like granted, upgraded, resold, void, withdrawn, evaluation,
transferred, etc., and then to produce signature S 2
S 2 Multihash Signature ( Hash (Image Portion 2802d) ， K pte0 )
(4) Vendor embeds S 2 as third watermark WM 2 to the bottom band of image
portion 2802c in blue using vendor's asymmetric WM private key K WM ， pte or
published symmetric WM key K WM2 to select pixel address locations for WM 2
embedding as in Figure 23， where K WM , pte or K WM2 may also act like stego-key
(5) Other remaining pixel locations in the blue band are filled with random bits
(6) Vendor debits the reseller's account for the sold software license
(7) Vendor records the licensee's information for this software license
(8) Vendor sends processed license SLC 2 to reseller and licensee via MePKC
V
Reseller or sales agent veriiying the processed digital software license SLC 2 :
(1) Reseller verifies WM 2 of CHQ 2 using vendor's asymmetric watermarking
public key K WM , pU b or published K WM2 ， and vendor's public key Kp Ub0
(2) If WM 2 is verified, reseller checks the account for the debit transaction
(3) Otherwise if WM 2 is rejected, reseller reports to the vendor for investigation
Licensee verifying the processed digital software license SLC 2 :
(1) Licensee verifies WM 2 of SLC 2 using vendor's asymmetric watermarking
public key K WM , pU b or published K WM2 ， and vendor's public key K pub0
(2) If WM 2 is verified, licensee checks one's licensing record at vendor's website
(3) Otherwise if WM 2 is rejected, licensee reports to the vendor for investigation
Figure 10.8 Endorsement process of a software license by a licensee in the triple-
watermark digital software license method and system
176
〜誦
〜蓮
/-^ 3002
/ 싀 扁
CHAPTER 11 APPLICATIONS OF BIG SECRET & MePKC
(PART 3)
11.1 MePKC Human-Computer and Human-Human Authentication Schemes
11.1.1 Related Works: Computer Password Authentication Protocol
In this networked info-computer age, computer-computer mutual
authentication uses asymmetric key cryptography, but human-computer and human-
human mutual authentications till now still stick to symmetric key cryptography. In
fact, the most frequently used application of secret is authentication access of a
human to a computer for online account access. The online computer authentication
methods (Beutelspacher, 1994) using password the secret include (i) simple
transmission of key, (uj transmission of encrypted key, (m) transmission of key
through encrypted channels, (iv) hash-based challenge-response method, (v) zero-
knowledge password proof, and (vi) PAKE (Password- Authenticated Key Exchange).
All of these six methods are based on a shared secret between a user and the server.
The first method using simple transmission of key in the clear channel is an
insecure approach. The second method using transmission of encrypted key is in fact
firstly proposed by H. Feistel (1974a, 1974b, 1974c) in his three patents, US Patents:
US3798359 "Block Cipher Cryptographic System", US3798360 "Step Code
Ciphering System", and US3798605 "Centralized Verification System", filed on the
same day on 30 June 1971. For the third method using transmission of key through
encrypted channels, the encrypted channels are based on the protocols like SSL
(Secure Sockets Layer) or TLS (Transport Layer Security). Later, hash function is
created and subsequently the fourth method called hash-based challenge-response
method using hashed password, where a server stores the hash value of a password.
The second, third, and fourth methods remain as the current most popular online
computer authentication methods till today.
For the fifth method called zero-knowledge password proof, it is more
complicated where a secret owner can prove to a verifier its ownership of a secret
without revealing the secret. The fifth method is somehow modified to become the
177
sixth method called PAKE. Examples of PAKE include EKE (Encrypted Key
Exchange), PAK (Password- Authenticated Key exchange), PPK (Password-
Protected Key exchange), SPEKE (Simple Password Exponential Key Exchange)
(Jablon, 2006)， SRP-6 (Simple Remote Password Protocol version 6) (Wu， 2003)， etc.
For a good computer password authentication protocol, there are three main
issues to be fulfilled: Resistance to dictionary attack, (perfect) forward secrecy, and
non-plaintext equivalence. Forward secrecy means resistance to compromise secret
information if another part of the protocol is compromised. Perfect forward secrecy
means the compromise of long-term key used to derive an agreed ephemeral key
does not compromise the agreed keys from earlier runs. Non-plaintext equivalence
means a data that cannot be used to gain the same access level of a key/password.
Computer password authentication protocols that can resist dictionary attack
are EKE (Encrypted Key Exchange) family of protocols and a few public-key
assisted protocols. Protocols that can fulfill the conditions of resistance to dictionary
attack and prefect forward secrecy are the strongest members of EKE family of
protocols like DH-EKE (Diffie-Hellman Encrypted Key Exchange) and SPEKE
(Simple Password Exponential Key Exchange). SPEKE was firstly proposed by D. P.
Jablon (2004) in US Patent: US7010692 "Cryptographic Methods for Remote
Authentication". For protocol that can fulfil all the three issues of resistance to
dictionary attack, prefect forward secrecy, and non-plaintext equivalence, there is
currently only one called SRP-6 (Simple Remote Password Protocol version 6). SRP
was firstly proposed by T. J. Wu (1998) in US Patent: US6539479 "System and
Method for Securely Logging onto a Remotely Located Computer".
Nevertheless, the PAKE of SRP-6 still has a long-term shared secret and is
not yet a fully asymmetric key crypto system. Hence, if the long-term shared secret is
re-used, SRP-6 is subject to malicious server attack, where the faulty server having
the username, salt, and verifier can pretend to be the another actual server using the
same secret. Moreover, it is lacking of mutual authentication. As compared with the
MePKC authentication methods and systems in the preferred embodiment of this
thesis, SRP-6 also has more rounds of message exchange, more IP packets and
longer processing time.
178
For authentication protocol operating on the platform of asymmetric key
crypto system, split private key crypto system has a few protocols for these purposes.
However, the private key of split private key crypto system is only partially
memorizable and another portion of private key is stored in the authentication server.
The weakness of split private key crypto system is a malicious authentication server
can launch guessing attack and dictionary attack over the first portion of
memorizable split private key. Hence, there exists a need to have a password
authentication protocol for human-computer and human- human interfaces that
operates on the asymmetric key crypto system using a fully memorizable private key
for each user.
11.1.2 First Model without Perfect Forward Secrecy
Yet in the seventh application 9.1(vii) of the present invention in applying the
created big memorizable secret, two MePKC human-computer and human- human
authentication schemes between a human user and a local computer or remote server
(or human user) over an insecure computer communication network are presented.
Challenge-response authentication protocol is adopted for these authentication
schemes without any shared secret and transmission of secret key over the insecure
channel. The challenge has a nonce to resist replay attack. Nonce stands for "number
used once" and may be a one-time random number, counter, or time stamp. Yet one
of many advantages is no storage of encrypted password, hashed password, verifier,
or shared secret in the local or remote computing system. Subsequently, this MePKC
authentication scheme can also resist phishing attack and spoofing attack that try to
steal user password.
Since there is no storage of password, system and network administrators will
no longer know the secret of any user's key. This allows a user to use the same
asymmetric key pair for different offline/online accounts. By sharing the same
asymmetric key pair among different accounts, the memorizability of a user is
improved, and hence there is no more need to jot down various keys in the notebook.
Since there is no encrypted password, hashed password, or verifier, the pre-
computation attack can be avoided. Other attacks such as guessing attack, dictionary
179
attack, and brute force attack will still be possible. However, guessing attack and
dictionary attack can be avoided if the 2D key, multilingual key, multi-tier geo-
image key, or multi-factor multimedia key is used properly as for the key style of
ASCII art and Unicode art. If the same asymmetric key pair is used together with
multihash key to create different slave keys for different online accounts, this allows
pseudo-one-set password entry to multiple websites without having password
domino cracking effect as in the symmetric key crypto systems.
However, the disadvantage of MePKC authentication schemes is the slow
processing speed of PKC. Hence, the size of challenge message has to be limited to
only a few units of encryption block of PKC, like block size of 256 to 512 bits for
256- to 512-bit MePKC, respectively. A wonderful authentication scheme over a
computer communication network shall have the features of non-plaintext
equivalence, prefect forward secrecy, and resistance to dictionary attack.
For the first basic model of the MePKC authentication scheme as in Figures
1 1.1-1 1.2， it has the features of non-plaintext equivalence internally and resistance to
dictionary attack externally by using secret creation method of 2D key, multilingual
key, multi-tier geo-image key, or multi-factor multimedia key. The first basic model
is still lacking of the feature of prefect forward secrecy, because the compromise of
long-term private key used to derive an agreed ephemeral key does compromise the
agreed keys from earlier runs.
11.1.3 Second Model with Perfect Forward Secrecy
To include the feature of prefect forward secrecy, the second model of
MePKC authentication scheme as in Figures 11.3-11.5 is innovated. Now, a human
user may use multihash key and has a long-term asymmetric key pair [K pte uL ， K pubU L]
and a one-time asymmetric key pair [K pte u ， K pub u] acting as rolling key for each
login or authentication access. Now, the compromise of long-term private key used
to derive an agreed ephemeral key does not compromise the agreed keys from earlier
runs. An added feature for this second model is the optional inclusion of a key
exchange scheme to establish a shared key between the human user and remote
server.
180
From 101
Creating a human user's private key with sufficient key entropy for n-bit MePKC:
(1) User U creates a big memorizable user's private key K pteU with entropy E K
from Box 101
(2) If E K < n, then go to 100 again to create another K pteU as in Box 101
(3) Else if E K > n, then generate user's public key K pubU using K pteU
Kp UbU Public Key Generation (K pte u)
/X 기
3100
Figure 丄 1.1a Creating a sufficiently big and yet memorizable user's private key
From 3100
V
New human user registering an offline/online account for authentication access:
(1) User U accesses a local computer system S L or remote server S R
(2) User creates and sends a username ID to computer S L or S R
(3) If the ID is unique and availaole, computer S L or S R accepts the ID and
requests for user's public key K pubU ; otherwise user creates another ID
(4) User sends K pubU to computer S L or S R for storage and future authentication
access
Figure 丄 1.1b Account registration of a new user
From 3204 or 3205
V
Human user U changing the registered public key K pubU to new public key K pubu ， :
(1) Once getting authentication access from Box 3204 or 3205， user can create a
new user's public key K pubU ' as in Box 3100
(2) User sends K pubU ' to the local computer S L or remote server S R to replace the
old user's public key K pubU for next login
Figure 11.1c Replacing a user's public key by a user
Figure 11.1 Various not- so-frequent operations of the basic model of MePKC
authentication schemes with feature of non-plaintext equivalence
181
3200
\レ
A registered human user U attempting to login to an offline/online account:
(1) User U accesses a local computer system S L or remote server S R
(2) User sends one's registered username ID to computer S L or S R
V
Computer S L or S R creating a challenge C for user to gain authentication access:
(1) Computer S L or S R creates a challenge C using an n-bit random bitstream B，
timestamp T， and a nonce N R
C ^ ( B II T II N R )
(2) Computer S L or S R encrypts the C using user's public key K pubU to produce C E
C E Public Key Encryption ( C ， K pubU )
(3) Computer S L or S R sends encrypted challenge C E to the user through SSL
V
User decrypting the encrypted challenge C E to get a response R:
(1) User decrypts the C E using user's private key K pteU to produce response R
R <— Private Key Decryption ( C E ， K pteU )
(2) User encrypts the R using public key K pubS of computer S L or server S R to
produce encrypted response R E
R E <— Public Key Encryption ( R ， K pubS )
(3) User sends encrypted response R E to the computer S L or S R through SSL
Computer S L or S R decrypting the encrypted response R E to verify user's access:
(1) Computer S L or S R decrypts R E using its private key K pteS to produce R
R <— Private Key Decryption ( R E ， K pteS )
(2) If R ^ C, the user's authentication access is rejected, and user's further action
is directed to 3202 for another authentication attempt based on some rules
(3) Otherwise if R = C， the user's authentication access is verified and granted
(4) Computer S L or S R informs the user that user's authentication is successful
V
For mutual authentication in a remote computer communication network, go to
3200， and invert the roles of human user and remote computer S R
Figure 11.2 rirst basic model of MePKC authentication scheme between a human
user and a computer with features of non-plaintext equivalence and optional mutual
authentication
182
〜遍
3202
/XT'
시기 通
〜通
3300
Human user holds a long-term private key K pte uL and published public key K pubUL .
New human user registering an offline/online account for authentication access:
(1) User U accesses a local computer system S L or remote server S R
(2) User creates and sends a username ID to computer S L or S R
(3) If the ID is unique and available, computer S L or S R accepts the ID and
requests for user's public key K pubU ; otherwise user creates another ID
N レ
Creating a human user's authentication private key K pteU with sufficient key
entropy for n-bit MePKC and user's authentication public key K pubU :
(1) User U creates a big memorizable user's secret key K P with entropy E P from
Box 101 and an n-bit salt s from a CSPRBG
(2) If E P < n, user goes to 100 again to create another K p as in Box 101
(3) Else if E K > n, user generates user's private key K pteU and public key K pubU
Kpteu Hash ( K P II ID II 5 ) , K pubU Public Key Generation (K pteU )
(4) User signs the K pubU using K pte uL to produce signature S pub K
SpubK Sign ( Kp U bU ， KpteUL )
(5) User sends K pubU , s, and S pub K to computer S L or S R for storage and future
authentication access
(6) Computer S L or S R stores K pubU in ciphertext, as well as s and S pub K in plaintext
Figure 11. Account registration of a new user by creating a sufficiently big and yet
memorizable user's private key
From 3500
V
Human user U changing the registered public key K pubU to new public key K pubu ， :
(1) After getting authentication access from Box 3500, user creates new salt s\
user's private key K pteU ' and user's public key K pubU ' as in Box 3302
K^u ' Hash ( K P II ID II ぶ，)， K pubU ， Public Key Generation (K ptelJ ，)
(2) User signs the K pubU ， using K pte uL to produce signature S pub K，
SpubK' ^~ Sign ( Kp U bU ' ， KpteUL )
(3) User sends K pubU ，， s\ and S pU bK， to the local computer S L or remote server S R
to replace the old authentication dataset K pubU , s, and S pub K
(4) Computer S L or S R stores K pubU ， in ciphertext, as well as s' and S pub K， in
plaintext for next login
Figure 11.3b Replacing a user's authentication dataset like user's public key and salt
by a user
Figure 11.3 Various not- so-frequent operations of the second model of MePKC
authentication schemes with features of non-plaintext equivalence and perfect
forward secrecy
〜遍
ᄉ\71 濯
183
3400
A registered human user U attempting to login to an offline/online account:
1) User accesses a local computer system Sl or remote server Sr
2) User sends one's registered username ID to computer S L or S R
Computer Sl or Sr creating a challenge C for user to gain authentication access:
1) Computer S L or S R looks up the corresponding K pubU , ん and S pub K of username ID
2) Computer S L or S R encrypts K pubU using K pubU to produce ciphertext CK pubU
CKp UbU Public Key Encryption ( K pubU ， K pubU )
3) Computer S L or S R creates and encrypts a challenge C using an n-bit random bitstream
B， timestamp T， and a nonce Nr
C — ( B II T II N R ) ， C E — Public Key Encryption ( C ， K pubU )
4) Computer S L or S R signs the concatenation of Si, CK pubU , and C E for integrity checking
using private key of computer or server K pteS to produce signature Ss
S s ― Sign ( Hash ( s, II CK pubU II C E )) ᄂ
5) Computer S L or S R sends S\, CK pubU , C E , and S s to the user through SSL
3401
3402
User decrypting the encrypted challenge Ce to get a response R and shared key K SH :
1) If Ss is rejected, go to 3400; else if Ss is verified, go to step (2) of Box 3403
2) User generates K pteU and then K pubU , and decrypts CK pubU to get K pubU2
K pt eu Hash ( K P II ID II め)， K,
-pubU '
Public Key Generation (K pteU )
K pubU2 ― Private Key Decryption ( CK pubU ， K pte u )
3) If K pubU + Kp Ub u2 ， go to 3400; else if K pubU = K pubU2 ， computer S L or server S R is
authenticated and go to step (4) of Box 3403
4) User decrypts the Ce using user's private key K pteU to produce response R
R <— Private Key Decryption ( Ce ， K pte u )
5) User creates a shared key K SH with server Sr by hashing R
R = ( B II T II N R ) , K SH ^- Hash (R) ' ᄂ
6) User encrypts the R using public key K pubS of computer Sl or server Sr to produce
encrypted response R E
R E Public Key Encryption ( R ， K pubS )
7) User creates new salt s 2 , user's private key K pteU2 ， and user's public key K pubU2 as in
Box 3302 '
3403
K,
-pteU2 "
Hash ( K P II ID \\s 2 ), K,
pubU2 "
Public Key Generation (K pteU2 )
8) User signs the K pubU2 using K pteU L to produce signature S pubK2
SpubK2 ᅳ Olgn ( K pubU 2 ， KpteUL )
9) User sends R E ，&, Kp UbU2 ， and S pub K2 to the computer S L or server S R through SSL
To 3500
Figure 丄 1.4 Second model of MePKc authentication scheme between a human user
and a computer with features of non-plaintext equivalence, perfect forward secrecy,
and optional key exchange scheme (Part 1)
184
3500
V
Computer S L or server S R decrypting the encrypted response R E to verify user's
access and to get a shared key Ks H :
(1) Computer S L or server S R hashes the K pubU2 to get hash value H m
Hui Hash (K pubU2 )
(2) Computer S L or server S R decrypts S pub K2 using K pubUL to get hash value H U2
H U2 Public Key Decryption (S pubK2 ， K pubUL )
(3) If Hui ^ H U2 ， Sp Ub K2 is rejected, and user's further action is directed to Box
3402 for another authentication attempt based on some rules; else if Hui =
H U2 ， S pub K2 is verified, go to step (4) of Box 3501
(4) Computer S L or S R decrypts R E using its private key K pteS to produce R
R Private Key Decryption ( R E ， K pteS )
(5) If R ^ C， the user's authentication access is rejected, and user's further action
is directed to Box 3402 for re-authentication attempt based on some rules
(6) Otherwise if R = C， the user's authentication access is verified and granted
(7) Server S R creates a shared key Ks H with human user U by hashing R
R = ( B II T II N R ) ， Ks H い Hash (R)
(8) Computer S L or server S R stores K pub u2 in ciphertext, as well as s 2 and S pub K2 in
plaintext for user's next login or authentication access
(9) Computer S L or S R informs the user U that user's authentication and/or key
exchange is successful
V
Human user U and remote server S R can use the shared key Ks H for any
application using secret over an insecure computer communications network
Figure 丄 1.5 Second model of MePKc authentication scheme between a human user
and a computer with features of non-plaintext equivalence, perfect forward secrecy,
and optional key exchange scheme (Part 2)
11.1.4 Re-Authentication Rules
Mutual human-computer authentication for both the first and second models
is possible, and it is also extendable to mutual human- human authentication over a
computer network. For railed authentication, there are some re-authentication rules
for another login attempt and so on. These re-authentication rules include limited
time, limited usage amount of a factor, limited number of allowable attempts per unit
of time, CAPTCHA activation, secret question(s) and answer(s), as well as password
throttling using time, bit length, and crypto system, etc.
〜舊
185
11.2 MePKC Digital Certificate Having More Than One Asymmetric Key
Pair for Different Protection Periods and Password Throttling
11.2.1 Related Works: Digital Certificate and Password Throttling
In using PKC (Public-Key Cryptography), a user needs to bind one's public
key with one's identity. The file binding the user's identity and public key is called
digital certificate (aka public-key certificate). Digital signature is used to bind the
user's identity and public key by an introducer using web of trust or by a trusted third
party (TTP) using certification authority (CA).
In the current prior art, there is only one public key per digital certificate. In
PKC, different key sizes correspondent to different protection periods. A short key
size like RSA-1024 will have to be changed or revoked frequently. Frequent
certificate revocation may cause complicated management problems. Hence, a
private key has to be steady throughout its validity period to avoid frequent
certificate revocation. Successful cracking of encrypted private key, as well as
forgetfulness of symmetric key encrypting the private key and partially memorizable
private key tend to fail this purpose. Therefore, the ciphertext of the encrypted
private key has to be hidden from the public domain.
For online account using split private key crypto system, attackers may launch
online dictionary attack to the server. The method of locking an account after a pre-
set number of unsuccessful login attempts is not practical because it is subject to
denial-of- service attack. The follow-up services to re- activate the account through
phone and face-to-face communications are tedious and costly. Consequently, split
private key crypto system was improved by Sandhu, deSa, and Ganesan (2005a) to
have the function of password throttling using the increasing complexity of time
response and bit length for unsuccessful authentication. The time response will be
slower or the bit length of the challenge will be longer whenever a previous login
attempt is unsuccessful until a maximum pre- set value tolerable by a user. A slight
modification is to measure based on limited number of login attempts per time unit.
The disadvantage of this method is that a digital certificate with short
asymmetric key pair like RSA-1024 will still have to be changed frequently. Another
186
disadvantage is that there is a maximum of time response and processing time like
one second that a user can tolerate. A delay of one second adds only by about 20 bits
on the platform of contemporary computing technologies. Yet in some password
generation systems, key strengthening (aka key stretching) is use to harden a
password by hashing a password seed for many rounds of iteration for a pre- set time
unit like one second to freeze the demand of better computing technologies for
longer key length. It tells that password throttling (Sandhu, deSa & Ganesan, 2005a)
using time response may be not tolerable if it is used together with key strengthening.
Hence, there exists a need to improve this method to have lower frequency of
certificate revocation and yet fast time response. Moreover, there is a need to have
bigger memorizable secret to resist online dictionary attack and malicious server
attack over the split private key crypto system.
Another method to resist machinery online dictionary attack is to use
CAPTCHA (Completely Automated Public Turing test to tell Computers and
Humans Apart) by asking a user to key in some data presented by a computer that
cannot be interpreted by another remotely networked computer trying to attack the
account. This method is quite effective but it cannot extend the validity of a digital
certificate with short asymmetric key pair like RSA-1024 that is still changed or
revoked frequently. Hence, need exists to extend the digital certificate validity to
reduce the certificate revocation frequency via a better password throttling method.
11.2.2 MePKC Digital Certificate with Many Asymmetric Key Pairs
Yet in the eighth application 9.1(vm; of the present invention in applying the
created big memorizable secret, the multihash key allows the usages of multiple
secrets for various applications and this can realize the MePKC digital certificate
having more than one asymmetric key pair. Due to technical security and legal
factors, a pair of asymmetric key cannot be re-used for different cryptographic
schemes like encryption, signature, and authentication. Hence, it is very common for
a user to own more than one asymmetric key pair. Here, MePKC digital certificate
with four public keys is illustrated in Figure 11.6 for one of its various functions
according to private key sizes, protection periods, and difficulty levels of cracking.
187
3600
V
Types of asymmetric key pair in an n-bit MePKC digital certificate having four public
keys for various applications, such as password throttling:
(1) 160-bit MePKC: 160-bit memorizable private key, or private key from a multi-factor
key of 80-bit memorizable secret and 160-bit software token
(2) 256-bit MePKC: 256-bit memorizable private key, or private key from a multi-factor
key of 128-bit memorizable secret and 256-bit software token
(3) 384-bit MePKC: 384-bit memorizable private key, or private key from a multi-factor
key of 192-bit memorizable secret and 384-bit software token
(4) 5 12-bit MePKC: 512-bit memorizable private key, or private key from a multi-factor
key of 256-bit memorizable secret and 512-bit software token
3601
Different n-bit asymmetric key pairs for different cryptographic applications based on
different protection periods or difficulty levels of cracking:
(1) 160-bit MePKC: 5 -year protection or till year 2010 or use key stretching to freeze the
quest for longer key length
(2) 256-bit MePKC: 30-year protection
(3) 384-bit MePKC: 150-year protection
(4) 5 12-bit MePKC: 300-year protection or resistance to future quantum computer attack
/ 싀
3602
\レ
Password throttling using different MePKC cryptosystems based on different difficulty
levels of cracking for re-authentication rules after failed login attempt as in Boxes 3204
and 3501 in MePKC authentication schemes:
(1) For the first 2 4 re-authentication attempts, 160-bit MePKC or higher level without
request for CAPTCHA
(2) For the second 2 6 re-authentication attempts, 160-bit MePKC or higher level with
request for CAPTCHA ᄂ
(3) For the third 2 6 re-authentication attempts, 256-bit MePKC or higher level with
request for CAPTCHA ᄂ
(4) For the fourth 2 re-authentication attempts, 384-bit MePKC or higher level with
request for CAPTCHA
(5) For the fifth 2 6 re-authentication attempts within a period t, 512-bit MePKC or higher
level with request for CAPTCHA
(6) If more than the fifth 2 6 re-authentication attempts within period t, resort to symmetric
key cryptosystem and secret Q&A sessions, or a phone/face-to-face authentication
(7) Otherwise if more than the fifth 2 6 re-authentication attempts and outside period t, go
to step (5) of Box 3603
(8) If a user succeeds in at least one re-authentication attempt, system access is granted
3603
Figure 丄 1.6 MePKC digital certificate with four public keys for various applications.
such as password throttling
188
The illustrated public key settings of a MePKC digital certificate are 160， 256.
384， and 512 bits, in which their private keys may be created from multi-factor
multilingual key. For re-authentication rules after failed login attempts, password
throttling based on crypto system is presented as one of its potential main functions.
Other password throttling techniques use different periods of response time and
lengths of challenge message. After series of password throttling, the authentication
scheme may resort to symmetric key crypto system and secret Q&A (Questions and
Answers) session for limited information access, or phone/face-to-face authentication
to re- activate the account. Another potential function is to let the MePKC digital
certificate to have at least a bait asymmetric key pair. This bait will detect if there is
any criminal crony interested with any MePKC digital certificate.
11.3 Three-Tier MePKC Digital Certificates for Ladder Authentication
11.3.1 Related Works: Digital Certificate and Ladder Authentication
For Internet banking using password the secret for authentication access,
usually more than one factor and one authentication process are needed for different
services due to the sensitiveness and critic ality of monetary matters. For instance, a
first symmetric key through computer communications network is needed to login to
an Internet banking account. A second random number the secret, that is sent from a
bank server to a user's mobile phone through another communication channel, is
needed to activate some financial services like fund transfer and utility bill payment,
as well as non-financial services like changes of mailing address, email, and phone
number. These different authentication processes for different sensitive services of an
account is called ladder authentication (Sandhu, deSa & Ganesan, 2006e).
Although this method is effective, it limits to users' with mobile phone and
the costs of SMS (Short Message Service) to deliver the random number can be quite
a large amount when the Internet banking is prevalent. For example in Malaysia,
there are a population of 27 million and an average household size of five members
per family in 2007. Let each household have five types of utility bills per month:
Water, electricity, one wired phone, and two wireless phones. Then, there are 27
189
million bills per month throughout Malaysia. If an SMS is charged one cent by the
services provider of mobile phone, then it is MYR$3.24 million annually.
The operating costs become higher if a mobile phone is registered overseas.
This is a norm of phenomenon for a Malaysian using Singapore Internet banking
services, and vice versa. To solve this problem in Singapore, where lots of its
residents are occasionally residing overseas, Singapore banks use the one-time-
password token (OTP token) like RSA SecurlD token. The seeded OTP token creates
temporary password with a finite usable life such as thirty seconds. For every cycle
of usable life, another temporary password is generated. An authentication server
knows the seed and each usable temporary password as well as its usable life, based
upon shared algorithms with the OTP token. An overseas user uses the temporary
password from the OTP token to replace the random number of an SMS.
Nevertheless, the OTP token is subject to loss, damage, and mobility
convenience. Bank will charge the users for replacement of an OTP token due to loss
or damage. Currently in Singapore, the replacement cost is SGD$20 per unit of OTP
token. Moreover, the temporary password of OTP token is displayed in plaintext
mode. Anyone who gets the OTP token can subsequently obtain the temporary
password. In a summary, in the current prior art, the ladder authentication methods
using SMS of mobile phone and OTP token incur a high operating cost. Hence, there
is a need to apply specific PKC digital certificate using fully memorizable private
key to implement a cost- saving and yet securer ladder authentication system.
11.3.2 Three-Tier MePKC Digital Certificates for Various Applications
In the ninth application 9.1(ix) of the present invention in applying the
created big memorizable secret, three-tier MePKC digital certificates can perform the
functions of persistent private key, rolling private key, and ladder authentication as in
Figure 11.7. The number of tier can also be other values depending on the design
requirements. The first group at the first tier acts as the introducer or endorser for the
other groups. The user information of the digital certificates in the second and third
groups can be updated easily from time to time.
190
3700
V
Group types of three-tier MePKC digital certificates for various applications,
such as persistent private key, rolling private key, and ladder authentication:
(1) First group at the first tier Gi : Acting as certification authority, introducer or
endorser of web of trust for the second and third groups of three-tier MePKC
digital certificate
(2) Second group at the second tier G 2 : Two subgroups for non-persistent and
persistent private keys with optional feature of rolling private key K R using the
update of salt
Kg2 <— K R <— Hash ( Master Key II Username ID II salt ) or
Kg2 ^ K R ^- Hash ( Multihash Key (Master Key II Username ID) ， salt )
(2.1) First subgroup of second group G 2 si : Non-persistent private key for
ephemeral or transient usages like one-time authentication
(2.2) Second subgroup of second group G 2 s2 : Persistent private key within
limited time, limited number, or limited number per time unit, for steady
usages like fund transfer
(2.2.1) Sub-subgroups of second subgroup of second group, G 2 s2si, G 2 s2S2, ''.，
G 2 s2Sn : For ladder authentication, where different sub-subgroups are given
rights to access, manage, modify, endorse, delete, etc., different set of info
(3) Third group at the third tier G 3 : For highest security level, where the private
key in this group is only created and used when the network access of the
computer is disconnected
(4) Each group may be digital certificate with one or more asymmetric key pairs
An example of using three-tier MePKC digital certificate in Internet banking:
(1) Use multihash key to create multiple memorizable private keys for different
groups of three-tier MePKC digital certificate
(2) The public key in Gi is signed by a trusted third party being a certification
authority or introducer of web of trust to become a digital certificate
(3) Private key in Gi is used to sign and endorse other public keys in the second
and third groups
(4) Private key in G 2 si is used for one-time authentication access to the website
(5) Private key in G 2S2S1 is used to access and manage first group of information
like changing personal particulars
(6) Private key in G 2 s2S2 is used to access and manage second group of
information like fund transfer
(7) Private key in G 2 s2s n is used to access and manage n-th group of information
(8) Private key in G 3 is used for highest security when network is disconnected
like fund transfer more than a preset amount to a third party
3702
Figure 1 1.7 Three-tier MePKC digital certificates for various applications, such as
persistent private key, rolling private key, and ladder authentication
191
The second group has two subgroups with the optional feature of rolling
private key, which means regular replacement of asymmetric key pair. Each rolling
private key is updated when the salt value is updated according to one of the two
equations, where the first equation is from the second model of the MePKC
authentication scheme as in Figures 11.3-11.5， and the second equation applies the
multihash key.
For the private key in the first subgroup of the second group, it is non-
persistent in computer memory for ephemeral or transient usages like one-time
authentication. For the private key in the second subgroup of the second group, it is
persistent in computer memory within limited time, limited number, or limited
number per time unit, for steady usages like changing personal particulars, fund
transfer and bill payment.
The second subgroup of second group can be further divided into many sub-
subgroups for ladder authentication to resist MITM (Man-In- The-Middle) attacks.
The private key in the first, second, third, ...， n-th sub-subgroups of the second
subgroup of the second group may be used to independently access, manage, modify,
endorse, delete, etc., first, second, third, ...， n-th groups of information, respectively.
The first and second groups can function to alternate and complement the
current prior art of authentication scheme in Internet banking, where first
authentication using password, and second authentication using SMS random
number or one-time-password token (OTP token). This SMS random number is
called specifically as TAC (Transaction Authorisation Code or Transaction
Authentication Code), TAP (Transaction Authorization Pin), Auth Code, and
Authorization Code in Internet banking as a second layer of protection. The ladder
authentication using different groups from different tiers of MePKC digital
certificate can be applied to Internet banking, as well as online share trading.
For highest security, the private key of the third group is only used when the
networked computer is offline or disconnected from the computer communications
network like Internet and LAN. When anonymity feature is needed, then at least an
additional set of MePKC digital certificate from the first, second, and/or third group
is needed.
192
11.4 Anti-Phishing Using MePKC Authentication Schemes
According to APWG (Anti-Phishing Working Group) (2008)， the numbers of
crimew are- spreading URLs infecting PCs with password-stealing code rose 93% in
2008 Ql to 6500 sites, nearly double the previous high of November, 2007， and an
increase of 337 percent from the number detected end of 2007 Ql. Patel and Luo
(2007) took a closer look at phishing, which is a serious crime for the Internet
banking (Hilley, 2006; Mannan & van Oorschot, 2007).
Lately, Microsoft has declared war against the phishing (Hunter, 2006) by
taking actions like recruiting the anti-phishing crusader (CastleCops, No date;
McMillan, 2008a). There is also some key management tools to resist the phishing
attack like Passpet (Yee & Sitaker, 2006).
Here, it is to note that when the seventh to ninth applications 9.1(vii)-9.1(ix)
of the present invention in applying the created big memorizable secret are applied
for authentication to access online accounts like Internet banking and online share
trading, the phishing attack can be thwarted easily since there are no shared secret at
all and no transmission of secret key between the human and computer over an
insecure computer communications network (Ford, 1994).
193
CHAPTER 12 APPLICATIONS OF BIG SECRET & MePKC
(PART 4)
12.1 Archiving the Voice/Video Calls of Wired/Wireless Phones
12.1.1 Related Works
Yet there is another important application of PKC (Public-Key Cryptography)
using fully big memorizaole secret. Here, the application of secret to mobile phone
(aka wireless phone, cellular phone, cell phone, and hand phone) (Lee, 1995) is
discussed. Since the invention of wireless telephone in the 1907 by Nathan B.
Stubblefield (1908) in the US Patent: 887357 "Wireless Telephone", filed on 5 April
1907， its number of functions keeps on increasing until now that even there is camera
capturing real-time image and making video call a reality.
One of the many inventions is by Charles A. Glaaden and Martin H.
Parelman (1979) entitled "Rapidly Deployable Emergencey Communication System"
and to introduce the concepts of frequency reuse and handoff. For mobile phone, it is
possible to record SMS, voice mail, local image and video. A user needs a passcode
(aka pin) the secret to access the voice mailbox. However, it is yet impossible to
download voice mail from a website and record interactive voice and video calls.
Moreover, the memory of mobile phone is limited due to its size and publicly
affordable selling price.
Nevertheless, there are commercial activities, legal cases, personal matters,
etc., that are constrained by physical distance and the most convenient
communications channel is a phone connection. Here, normally a wired phone will
be used together with a recorder to keep a copy of the conversation contents as
electronic evidence. However, having every household to own a phone recorder is
not cost-effective.
Hence, there exist needs to download voice mail from a website, as well as to
record, encrypt, store, access, manage, copy, download, and decrypt the interactive
voice and video calls from a website as electronic evidence. Distributed servers
located in the CO (Central Office) (aka telephone exchange) of wired phone and
194
MTSO (Mobile Telephone Switching Office) (DeVaney, Harper & Short, 1987) of
wireless phone shall be fully utilized for recording storage of voice and video calls.
Computer password authentication protocol using symmetric key crypto system, PKC,
or MePKC shall be used to access, manage, and download the recorded voice mail,
voice and video calls.
12.1.2 The Proposed Model
In the tenth application 9.1(x) of the present invention in applying the created
big memorizaole secret, MePKC authentication scheme is used to access a user
online account storing the recorded data like voice mail, voice call, and video call of
wired phone (aka wireline phone) and wireless phone (aka handphone, mobile phone,
wireless phone, cellular phone, cell phone) as in Figure 12.1.
A user's handphone has two buttons to select the call modes. For calling user,
if a first button is pressed, then a voice/ video session will be recorded and stored at
the distributed server. For called user, if the first button is pressed, the voice/video
call will be diverted to recording mode directly without receiving the call. Otherwise
if second button is pressed, the voice/video call of called user is received and there is
interaction between the calling and called users.
After the second button has been pressed, if the first button of called user is
not pressed until the end of a call, then no data will be recorded. Otherwise if the first
button of called user is pressed after the second button has been pressed, then the
following communicated data like voice, image, and video is recorded, encrypted,
and stored. Yet calling and called users may press the third and fourth buttons
accordingly to pause or terminate a recording session.
The distributed servers at the CO (Central Office) of PSTN (Public Switched
Telephone Network) of wired phone and/or CM (Communication Management) of
MTSO (Mobile Telecommunications Switching Office) of wireless phone records,
encrypts using MePKC, and stores the communicated voice/video call between the
calling and called parties. The voice/ video data is named, encrypted using MePK ᄂ，
and saved into the user account.
195
3800
\レ
Method and system to record, encrypt, and store the voice mail, voice call, and
video call in the distributed servers at the CO (Central Office) of PSTN (Public
Switched Telephone Network) of wired phone (aka wireline phone) and/or CM
(Communication Management) of MTSO (Mobile Telecommunications
Switching Office) of wireless phone (aka mobile phone, cellular phone):
(1) Calling user Ui may press a first button to record the voice/video session
(2) When called user U 2 receives a voice/video call, U 2 presses 1 of 2 buttons:
- First button to divert the call for recording storage without receiving the call
- Second button to receive the call without recording storage
(3) If first button is pressed, the distributed servers at the CO of wireline phone
and/or CM of wireless phone record, encrypt, and store call data Di
(4) Data Di is named, encrypted, and stored using MePKC into user LPs account;
(5) Else if second button is pressed, the user U 2 may later press the first button to
record the voice/video call
(6) If first button is not pressed after the second button has been pressed until the
end of the voice/video call, then no data will be recorded and stored;
(7) Else if first button is pressed after the second button has been pressed before
the end of the voice/video call, then distributed servers at CO of wireline
phone and/or CM of wireless phone will record and store the communicated
call data D 2
(8) Users Ui and U 2 may press the third and fourth buttons accordingly to pause
or terminate a recording session
(9) Data D 2 is named, encrypted, and stored using MePKC into user LPs account
3801
Method and system to access, download, and decrypt the recorded and stored data
of voice mail, voice call, and video call from the distributed servers at the CO
(Central Office) of PSTN (Public Switched Telephone Network) of wireline
phone and/or CM (Communication Management) of MTSO (Mobile
Telecommunications Switching Office) of wireless phone:
(1) User Ui or U 2 surfs the Internet website of the wired phone or wireless phone
services provider
(2) User authenticates oneself to access one's account in the distributed server at
CO of wireline phone and/or CM of wireless phone using any authentication
scheme like MePKC authentication scheme, SRP-6, etc.
(3) User searches and manages one's recorded data, Di and/or D 2 ， like voice mail,
voice call and video call
(4) User downloads selected data, Di and/or D 2 ， then decrypts at local computer
(5) User may select to subscribe to larger storehouse by paying more
(6) User logouts after all the transactions have been done
Figure 12.1 Operations to record, store, access, manage, and download the voice mail,
voice call, and video call in the distributed servers at the CO of PSTN of wireline
phone and/or CM of MTSO of wireless phone
196
The user can then surr the website of the wired phone and wireless phone
services provider to access one's account using MePKC authentication scheme or
other methods. Upon gaining access to the user account, the user may be optionally
required to gain a MePKC ladder authentication to further manage and download the
recorded and stored voice mail, voice call, and video call.
After downloading the encrypted data to a local computer, the user can
decrypt the data using MePKC schemes like hybrid encryption scheme of PKC and
symmetric key cryptography, where a symmetric key used to encrypt the voice/video
call is encrypted by a public key. Likewise, this method can be extended to other
online electronic data storage using MePKC authentication scheme.
12.2 Multipartite Electronic Commerce Transactions Using MePKC
12.2.1 Related Works
And yet there is crucial crypto system using secret to be improved soonest
possible. This crypto system is the current prevalent electronic commerce (aka e-
commerce) transactions (Kini & Choobineh, 1998; Konrad, Fuchs & Barthel, 1999;
Ahuja, 2000; H. M. Deitel, P. J. Deitel & Nieto, 2001; Ford & Baum， 2001; Kang,
Park & Koo, 2003; Lee, 2006a). In the current prior art, the electronic commerce
transactions operate in series of bipartite communication mode using credit card and
password the secret.
Once a user has selected a list of products to be purchased online at a certain
website (H. M. Deitel, P. J. Deitel & Nieto, 2000)， normally a credit card, such like
MasterCard or VISA, is then used to pay the bill, by sending the credit card number
and an optional secure code benind the card to the online merchant. For more
security, password the secret protecting the credit card may be requested by some
merchants. Examples of the services providers of credit card password are PayPal,
MasterCard SecureCode, and Verified by VISA. Graefe, Lashley, Guimaraes,
Guodabia, Gupta, Henry, and Austin (2007) discussed on the credit card transaction
security.
197
Besides merchant and credit card verifier for password, sometimes there
exists online loyalty point website demanding for another password authentication.
Hence, there are at least three rounds of bipartite communications for different stages
of authentication. In fact, a comprehensive electronic commerce transaction involves
many other entities such as merchant's bank, customer's bank, insurance company,
various departments of local, state, and federal governments, transportation agent,
storehouse agent, and so on. Each of this entity is now either usually paired with
merchant or rarely customer to one round of bipartite communication to initiate and
endorse a sub-process of an electronic commerce transaction.
Here, it can be observed that every individual round of bipartite
communications using token of credit card number and/or secret of a symmetric key
is not so secure and effective. It is in fact quite redundant and time-wasting. The
nature of an electronic commerce transaction is in fact a multipartite communication.
In dealing with cryptography and multipartite communications, there is a
branch of knowledge called BGP (Byzantine Generals Problem) (Lamport, Shostak
& Pease, 1982; Fischer & Lynch, 1982; Lamport, 1983; Aguilera & Toueg, 1999;
Pathak & Iftode, 2006). BGP involves a group of entities where loyal entities have to
reach a common agreement called BA (Byzantine Agreement) (Pease, Shostak &
Lamport, 1980; Yan & Chin, 1988; Meyer & Pradhan, 1991; Yan, Chin & Wang,
1992; Siu， Chin & Yang, 1998a, 1998b; Yan, Wang & Chin, 1999; Yan & Wang,
2005b; Fitzi & Hirt, 2006) at the end of a sufficient round of message exchanges,
regardless of the malicious and arbitrary messages communicated by faulty entities.
The solution of BGP is known as BAP (Byzantine Agreement Protocol), in
which BA can be successfully achieved based on the provided functions of PKC
(Public-Key Cryptography) like access control, authentication, non-repudiation, and
integrity. However, PKC popularity has to be boosted up by using fully big
memorizable secret to realize the MePKC. Faulty node detection and identification
are also needed (Wang & Yan， 2000; Yan & Wang, 2005a).
There are various types of available BAP. For the entities of electronic
commerce, they can be basically partitioned into three groups: Essential, government,
and non-essential groups. Here, there is a BAP also optimally divides a network of
198
entities into three partitions. This specific BAP is called tripartite ANN based BAP
(Tripartite Artificial Neural Network Based BAP) (aka Tripartite BAP-ANN or
Tripartite BAP with ANN) (Lee & Ewe， 2003) and developed from ANN based BAP
(Wang & Kao， 2001; Lee & Ewe， 2001， 2002， 2007b, 2007c; Lee, 2003).
The ANN here functions as a classifier and provides majority function over
rows and columns of MEM (Message Exchange Matrix) formed from three message
exchange rounds of Byzantine communications. For more details of ANN based
BAP and tripartite ANN based BAP, please refer to a master's thesis published on 25
October 2002 at Multimedia University, Malaysia, entitled "Artificial Neural
Network Based Byzantine Agreement Protocol" by Kok-Wah Lee @ Xpree Jinhua
Li (2003).
For the functioning, implementation, and optimization of ANN (Artificial
Neural Network), these literatures (Jacobs, 1988; Nguyen & Widrow, 1989, 1990;
Cooke & Lebby, 1998; Haykin, 1999) can be referred.
Again to emphasize here, e-commerce transaction involves multipartite
communications by nature and not many rounds of bipartite communications. The
BGP can model this multipartite cryptography problem of electronic commerce. BAP
is the solution of BGP, and hence multipartite communications of electronic
commerce. Tripartite ANN based BAP is well- suited to a network of e-commerce
entities divided into three groups.
Hence, there exists a need to realize e-commerce transaction based on
multipartite communications of BGP and BAP using MePKC, wherein the main
purposes are to speed up the processing time from many rounds of bipartite
communications and to rely on stronger security protection than the current prior art
using symmetric key cryptography.
12.2.2 Artificial Neural Network Based Byzantine Agreement Protocol (ANN
Based BAP)
In the eleventh application 9.1(xi; of the present invention in applying the
created big memorizable secret, MePKC cryptographic schemes like encryption and
199
signature schemes are used in the method and system of multipartite electronic
commerce (aka e-commerce) transactions using tripartite ANN based BAP (Artificial
Neural Network Based Byzantine Agreement Protocol) (aka tripartite BAP-ANN
(Tripartite BAP with ANN)) as in Figures 12.2-12.7(39-44) and article "Faulty Node
Detection in the Tripartite ANN based BAP" by Kok-Wah Lee and Hong-Tat Ewe
(2003). The MePKC provides the security like confidentiality, integrity,
authentication, access control, and non-repudiation to the tripartite ANN based BAP.
Other BAP can also be used for the multipartite e-commerce transactions.
Figure 12.2a shows the operating stages of a basic ANN based BAP. Figures
12.2b- 12.2c show the FCN (Fully Connected network) model and ANN architecture
for 4-node distributed network. The number of entities involved in the e-commerce
ranges from 4 to more than 30.
The simplest network of an e-commerce model includes merchant, customer,
bank, and a credit card company. For a big e-commerce model, it can be observed
that the partitioning of the large network into a few groups for k-partite ANN based
BAP is more efficient. This is because the bottleneck of processing time is the
number of exchanged messages that needs to undergo the MePKC encryption,
decryption, signing, and verifying processes. It is well-known that the operating time
of PKC is so slow that it is 1000 times slower than the symmetric key crypto system.
12.2.3 Entity Partitioning in the Electronic Commerce Transactions
From Figures 12.3a- 12.3b and 12.4b, it is known that tripartite partitioning is
the optimal k-partite ANN based BAP. Figure 12.4a shows the way to partition a
network into three partitions. Furthermore, from Figure 12.5， it is shown that the e-
commerce entities can be basically divided into three groups: Essential group,
government group, and non-essential group. For the first group, the entities of
merchant and customer are critical and cannot be replaced; whereas other entities are
non-critical and can be replaced. For the second group, all the entities are critical and
cannot be replaced. For the third group, all the entities are non-critical and can be
replaced. The source node now is the customer to confirm or cancel a buy order.
200
Figure 12.2a Block diagram of ANN based BAP
Figure 12.2b FCN model of 4-node distributed network
Figure 12.2c ANN model of 4-node distributed network
Figure 12.2 ANN based BAP and its smallest model of 4-node distributed network
201
■mio
Traditional BAP
ANN based BAP
4000
4 6 8 10 12 14 16 18 20 22
Number of processors, n
Figure 12.3a Traditional BAP and basic ANN based BAP
10 4
10 3
10 2
10 1
ANN based BAP with no partition
ANN based BAP with 3 partitions
/X 기
4001
10 15 20 25
Number of processors, n
30
Figure 12.3b Basic ANN based BAP and tripartite ANN based BAP
Figure 12.3 Total number of exchanged messages for different types of BAP
9876543210
〕 0000000000
5
ぷ ま will P5^lmlp,5s .Jo Jwclulnu {so TL
202
4100
Group 2
Group 1
O Node
Q Source
Image of source node
U3 Trusted party
Group 3 W O
Figure 12.4a Partitioning of a 10-node distributed network into three groups
10 4
10 3
10 2
つ^ 一
Ms g 匪 for tn 으
Msg mj/J for m=8
^ ^ ― ― ―
Msg 舰 for m=6
어- "
Msg 匪 for m=4
Msg mUi for m=4
Ms S llia Jor m=2
Msg, . for m=2
2 4 6 8
ANN based BAP with ^ partitions, k
10
4101
Figure 12.4b Optimal selection of network partitioning for tripartite ANN based BAP
Figure 12.4 Partitioning of a distributed network and its optimal partitioning
selection
203
First Group: Essential Group
(1) Merchant
(2) Customer
(3) Merchant's bank
(4) Customer's bank
(5) Credit card company
(6) Credit card password company
(7) Loyalty point company
(8) Local insurance company
(9) Foreign product-origin insurance company
(10) Foreign intermediate-region insurance company
/X 기
4200
4201
Second Group: Government Group
(1) National federal government (various departments)
(2) National state government (various departments)
(3) National local government (various departments)
(4) Foreign product-origin federal government (various departments)
(5) Foreign product-origin state government (various departments)
(6) Foreign product-origin local government (various departments)
(7) Foreign intermediate-region federal government (various departments)
(8) Foreign intermediate-region state government (various departments)
(9) Foreign intermediate-region local government (various departments)
4202
Third Group: Non-Essential Group
(1) Local land transportation agent
(2) Local air transportation agent
(3) Local sea transportation agent
(4) International foreign product-ongin land transportation agent
(5) International foreign product-ongin air transportation agent
(6) International foreign product-ongin sea transportation agent
(7) International foreign intermediate-region land transportation agent
(8) International foreign intermediate-region air transportation agent
(9) International foreign intermediate-region sea transportation agent
(10) Local storehouse agent
(11) Foreign product-origin storehouse agent
(12) Foreign intermediate-region storehouse agent
Figure 12.5 Partitioning of the entities involved in the electronic commerce
transactions into three groups: Essential group, government group, and non-essential
group
204
4300
Tripartite ANN based BAP for the multipartite communications of online
electronic commerce transaction to achieve a consensus or Byzantine agreement:
(1) Loyal message means customer decides to confirm the buy order
(2) Faulty message means customer decides to cancel the buy order
4301
Enter the initialization stage of tripartite ANN based BAP
4302
Simultaneously enter the message exchange stage and application stage of
tripartite ANN based BAP using MePKC for communications:
First round:
Each group applies basic ANN based BAP to achieve a group BA， A G .
Second round:
Each trusted party decides group B A，
A G , from each node in her own group.
Faulty node detection (FND) round:
Each node sends individual group BA，
A I; to other nodes in the other groups.
Third round:
Each trusted party interchanges group B A to decide a network B A, A N
Fourth round:
Each trusted party sends A G and A N to the nodes in her own groups.
Fifth round:
Each node compares the network BA， A N ， with individual group BA of each
node, A I; from the FND round to identify the faulty node(s) in the other groups.
4303
Enter the compromise stage of tripartite ANN based BAP to decide finally:
(1) Each node sends its Aj to customer the source node and customer derives A N
(2) If network BA is to confirm the buy order but faulty node exists in the non-
essential group, or essential group other than customer and merchant, go to 4300;
(3) Else if network BA is to confirm the buy order but faulty node exists in the
essential group for customer or merchant only, or government group, cancel the
buy order and exit;
(4) Else if network BA is to confirm the buy order and no faulty node, execute the
customer order to buy;
(5) Else if the customer decides to cancel the buy order, exit.
4304
Figure 12.6 Tripartite ANN based BAP with trusted party and faulty node detection
for multipartite electronic commerce transaction using MePKC cryptographic
schemes for communications
205
4400
Tripartite ANN based BAP for the multipartite communications of online
electronic commerce transaction to achieve a consensus or Byzantine agreement:
(1) Loyal message means customer decides to confirm the buy order
(2) Faulty message means customer decides to cancel the buy order
4401
Enter the initialization stage of tripartite ANN based BAP
4402
Simultaneously enter the message exchange stage and application stage of
tripartite ANN based BAP using MePKC for communications:
First round:
Each group applies basic ANN based BAP to achieve a group BA， A G ， and
detect the faulty node(s) inside the group.
Second round:
Each node sends her individual group BA， A I; to all the other nodes in the other
groups.
Third round:
Each node uses majority function over the received Aj from all the nodes in the
other groups to decide the A G of other groups. Then, each node decides the
network BA， A N ， from the three group BA.
Fourth round:
Each node compares A N with Aj from each node in the other groups to identify
the faulty node(s) in the other groups.
4403
V
Enter the compromise stage of tripartite ANN based BAP to decide finally:
(1) Each node sends its Aj to customer the source node and customer derives A N
(2) If network BA is to confirm the buy order but faulty node exists in the non-
essential group, or essential group other than customer and merchant, go to 4400;
(3) Else if network BA is to confirm the buy order but faulty node exists in the
essential group for customer or merchant only, or government group, cancel the
buy order and exit;
(4) Else if network BA is to confirm the buy order and no faulty node, execute the
customer order to buy;
(5) Else if the customer decides to cancel the buy order, exit.
Figure 12.7 Tripartite ANN based BAP without trusted party but still with faulty
node detection for multipartite electronic commerce transaction using MePKC
cryptographic schemes for communications
206
12.2.4 Tripartite ANN Based BAP with Trusted Party
Figure 12.6 shows a first implementation example of using BAP for the
multipartite e-commerce transaction having customer as the only source node.
Individual group BA， Ai, of each node equals to group BA， A G ， for loyal nodes but
not faulty nodes. Yet in a second implementation, both customer and merchant can
be source nodes for two independent Byzantine communications of e-commerce,
where one is the customer confirming the money payment for the buy order, and
another one is the merchant confirming the product/service delivery for the buy order.
For a very brief introduction, please refer to Lee (2006a).
12.2.5 Tripartite ANN Based BAP without Trusted Party
And yet in another third implementation as in Figure 12.7， the trusted parties
can be excluded if the individual group BA of each node is broadcasted to the nodes
of other groups and used directly to derive the network BA.
12.3 Trust Boosting of MePKC Digital Certificate by Using More than One
Certification Authority and/or Introducer of Trust of Web
12.3.1 Related Works: Risks of Public Key Infrastructure
The applications of PKI (Public Key Infrastructure) (Kuhn, Hu， Polk &
Chang, 200 1 ) in healthcare, finance, government, communications, etc., are
presented by Kapil Raina (2003). Meanwhile, for the applications of PKI in the
Internet protocols, one can refer to a book "Cryptography and Public Key
Infrastructure on the Internet" by Klaus Schmeh (2001). For the details operations on
how a user applies for a digital certificate through a CA (Certification Authority),
one can refer to a book "PKI: Implementing and Managing E-Security" by Andrew
Nash, William Duane, Celia Joseph, and Derek Brink (2001). It can be observed in
the third book that in the current prior, the CA generates the asymmetric key pair for
the user. This is not good because it may have malicious CA attack.
Yet Carl Ellison and Bruce Schneier (2000) discussed 10 PKI risks in their
article "Ten Risks of PKI: What You're not Being Told about Public Key
207
Infrastructure". The first risk on "Who do we trust, and for what?" questions on how
well the CA maintains its private keys well. The current digital certificate having
only one digital signature to certify its authenticity is not having a strong enough
trust. The successful cracking of a CA private key or existence of malicious CA
remains as a PKI risk. Corell (2000) had some comments on these PKI risks that
smartcard should be in favour.
The third risk on "How secure is the verifying computer?" questions on the
possibility of attacker adding its own public key to the list of certificate verification.
Again, the current digital certificate having only one digital signature to certify its
authenticity is not having a strong enough trust. The sixth risk on "Is the user part of
the security design?" questions on the degree of user involvement in the PKI. So far,
the user role is not strong in keeping one's secret because the asymmetric key pair is
still generated by the CA. A user holds only a symmetric key protecting the private
key of the asymmetric key pair. Hence, there exists a need to innovate the PKI to
allow user to create asymmetric key pair oneself and boost up the trust level of PKI.
The identity-related crime conspired by an organized crime group is getting
serious in today electronically networked info -computer age. One may refer to
UNODC (United Nations Office on Drugs and Crime) website to know more about
this identity-related crime at [URL: http://www.unodc.org/unodc/en/organized-
crime/index. html] (Identity Theft Resource Center (ITRC), No date; "Identity-
Related Crime," 2007; Wikipedia Contributors, 2008bh).
Some human interaction models are needed to simulate the group efficiency
of the organized crime group to fake the digital certificate. From the simulation, one
can design PKI that can make the organized crime group (Lampe, No date; Glick,
1995; Livingston, 1996; "UNODC and Organized Crime," No date; Layman &
Potter, 1997; Maxim & Whitehead, 1998; Chen, 2004; Ruan & Wang, 2005; Siegel,
2005; Wang, 2007; "Identity-Related Crime," 2007; He, 2007; United Nations Office
on Drugs and Crime (UNODC), 2004， 2008; Wikipedia Contributors, 2008s, 2008af)
to be inefficient and hence the PKI trust level can be increased.
Another approach is to enact warning punishment (Weiss & South, 1998;
Wikipedia Contributors, 2007r). Bierly, Kolodinsky, and Charette (No date)
208
discussed the relationships of creativity and ethical ideologies. Chatterjee (No date)
analyzed the equity ownership and the directors' board attitude for responsibilities,
obligations, and crimes. Khatri and Tsang (2003) surveyed and analyzed the
antecedents and consequences of cronyism (Wikipedia Contributors, 200 8q) in
organizations, which helped to know the malice probability of one or more CAs.
Kaneyuki Kurokawa (1997) proposed some very interesting and good human
interaction models in his paper entitled "Modeling Human Interactions". The studied
models are committee meeting, labour division, exploratory group, and technology
transfer. This article has somehow showed the coefficient of inefficiency of
Parkinson's Law by Professor Cyril Northcote Parkinson (1958， 2002)， in his book
"Parkinson's Law: Or the Pursuit of Progress". The coefficient of inefficiency
(Wikipedia Contributors, 2008i) ranges from 20 to 22 or more to trigger the
phenomena that a human group starts to become inefficient. Hence, there exists a
need to apply the results of these human interaction models over the organized crime
group to fake digital certificate in order to boost up the trust level of the digital
certificate.
There are other Kurokawa' s articles on human interaction models (Kurokawa,
1988， 1990， 1991). The Parkinson's Law has also been studied by some other
researchers (Aronson & Gerard, 1966; Aronson & Landy, 1967; Landy, McCue &
Aronson, 1969; "The Proof of Parkinson," 1969; Wikipedia Contributors, 2008x).
Klimek, Hanel, and Thurner (2008) analyzed the efficiency of a ministers' cabinet to
show the Parkinson's Law.
12.3.2 Some Human Interaction Models
In the twelfth application 9.1(xii) of the present invention in applying the
created big memorizable secret, method and system to boost up the trust level of
MePKC digital certificate by using more than one certification authority (CA) and/or
introducer of trust of web is designed. When one refers to the Figures 11.6-11.7 for
the MePKC digital certificate, one wih know that the private key and public key of a
user's asymmetric key pair is generated by the user and not the CA. This step can
avoid the malicious CA attack by giving the user to fully control one's private key
209
secret, and hence alleviating the sixth risk of Carl Ellison and Bruce Schneier (2000)
on "Is the user part of the security design?" questioning on the degree of user
involvement in the PKI.
For the first group of the user's asymmetric key pair of the three-tier MePKC
digital certificate as in Figure 11.7， it acts as the introducer of trust of web to the
other groups at tiers 2 and 3. For the certification of the first group instead, the
current prior art uses a single digital signature from a CA or introducer of trust of
web. However when the MePKC prevails, this prior art is not that appropriate in
view of the high demand of trust for the first group of three-tier MePKC digital
certificate. Innovated approach has to use to build up stronger trust by failing the
organized crime to fake MePKC digital certificate.
The possibility that the asymmetric key can be generated by a user allows the
user to bind one's identity, public key, and other data, into a binding file oneself. A
user can then request one or more CA and/or introducer of trust of web to sign,
certify, and issue digital signature. Every pair of binding file and a C A/introducer's
digital signature acts as a MePKC digital signature. Due to the independent trust of
each pair, other users only accept a binding file when all the pairs are verified.
Whenever there is one pair fails to be verified, then the user's binding file is rejected.
Hence, the more pair is the MePKC digital certificate, the lower is the probability to
successfully fake the user's MePKC digital certificate, the harder is the organized
crime group to be efficient, and the higher is the trust level of the user's first group of
MePKC digital certificate.
Coming to here, the Kaneyuki Kurokawa's human interaction models are
used to simulate the organized crime group to fake MePKC digital certificate.
Organized crime group has at least three persons to conspire a crime. Figure 12.8
illustrates the group efficiency of committee meeting. Figure 12.9 illustrates the
group efficiency of exploratory group. Figure 12.10 illustrates the success probability
of technology transfer. The models in Figures 12.8-12.10 are all developed by
Kurokawa and they are used in this article to derive Figures 12.11-12.13.
Kurokawa's model on committee meeting agrees with the coefficient of inefficiency
of Parkinson's Law ranging from 20 to 22 or more.
210
4500
Group Efficiency of Committee Meeting, GEc = n * p A (n-l)
n = Network size of human group
p = Probability of the chemistry being good between the chairperson and a member
n\p
0.500
0.700
0.850
0.900
0.000
0.000
0.000
0.000
1
1.000
1.000
1.000
1.000
2
1.000
1.400
1.700
1.800
3
0.750
1.470
2.168
2.430
4
0.500
1.372
2.457
2.916
5
0.313
1.201
2.610
3.281
6
0.188
1.008
2.662
3.543
0.109
0.824
2.640
3.720
8
0.063
0.659
2.565
3.826
9
0.035
0.519
2.452
3.874
10
0.020
0.404
2.316
3.874
4501
4.000
3.500
3.000
Group Efficiency of Committee Meeting
10 15
Network size of human group
令"
0.50
ᄀ뇨 p =
0.70
-©-p =
0.85
0.90
25
Figure 12.8 Group efficiency of a committee meeting according to the Kurokawa's
human interaction model
(l!un UBEnLO AoualoEm dnpIQ
211
4600
Group Efficiency of Exploratory Group, GEe = n * q A (n*(n-l)/2)
n = Network size of human group
q = Probability of the chemistry being good between a pair of members
n\ q
0.500
0.700
0.850
0.900
0.000
0.000
0.000
0.000
1
1.000
1.000
1.000
1.000
2
1.000
1.400
1.700
1.800
3
0.375
1.029
1.842
2.187
4
0.063
0.471
1.509
2.126
5
0.005
0.141
0.984
1.743
6
0.000
0.028
0.524
1.235
0.000
0.004
0.231
0.766
8
0.000
0.000
0.084
0.419
9
0.000
0.000
0.026
0.203
10
0.000
0.000
0.007
0.087
2.400
2.000
1.600
1.200
0.800
0.400
0.000
Group Efficiency of Exploratory Group
3 4 5 6
Network size of human group
4601
= 0.50
ᄀ뇨 q
= 0.70
夺 q
= 0.85
= 0.90
10
Figure 12.9 Uroup efficiency of an exploratory group according to the Kurokawa's
human interaction model
(llun ueEnLI) A0U3I0E 山 dnplo
212
4700
Success Probability of Technology Transfer, SP T = (p A (m-l+n)) * (q A n)
m = Number of ranks in the hierarchy, n = Number of receiving division, q = Probability of
the chemistry being good between a pair of peer members, p = Probability of the chemistry
being good between the chairperson and a member in a committee meeting
Curve
P
Q
R
S
n
1
l
3
3
P
0.750
0.850
0.750
0.850
m\ q
0.750
0.800
0.750
0.800
1
0.000
0.000
2
0.422
0.578
3
0.316
0.491
0.000
0.000
4
0.237
0.418
0.075
0.193
0.178
0.355
0.056
0.164
6
0.133
0.302
0.042
0.140
0.100
0.256
0.032
0.119
8
0.075
0.218
0.024
0.101
9
0.056
0.185
0.018
0.086
10
0.042
0.157
0.013
0.073
4700
Success Probability of Technology Transfer
Number of ranks in the hierarchy
Figure 12.10 Success probability of technology transfer according to the Kurokawa's
human interaction model
J3isue.ll AD0I0ULI031 Irussaoons Allllcleclpld
213
4800
Group Efficiency of Exploratory Group Formed from Leaders of Some Committee Meetings
(without condition for common consensus), GE EC0
tor m = 0, GE = 0; tor m = 1, GE = n*p A (n-l); tor m > 1， GE = ((n*p A (n-l))*m) + (m * q A (m * (m- 1 )/2))
m = Network size of human group of exploratory leaders, n = Network size of every committee
meeting, q = Probability of the chemistry being good between a pair of leader members, p =
Probability of the chemistry being good between the chairperson and a member in a committee
meeting
n
1
2
3
4
5
P
0.850
0.850
0.850
0.850
0.850
m\ q
0.800
0.800
0.800
0.800
0.800
0.000
0.000
0.000
0.000
0.000
1
1.000
1.700
2.168
2.457
2.610
2
3.600
5.000
5.935
6.513
6.820
3
4.536
6.636
8.039
8.906
9.366
4
5.049
7.849
9.719
10.875
11.489
5.537
9.037
11.374
12.819
13.587
6
6.211
10.411
13.216
14.950
15.871
7.065
11.965
15.237
17.260
18.335
8
8.015
13.615
17.355
19.667
20.896
9
9.003
15.303
19.510
22.111
23.493
10
10.000
17.000
21.675
24.565
26.101
4801
Group Efficiency of Exploratory Group Formed from Leaders of Some
Committee Meetings (without condition for common consensus)
28.000
24.000
23456789
Network size of human group of exploratory leaders
Figure 12.11 Group efficiency of an exploratory group formed from leaders of some
committee meetings (without condition for common consensus) as modified and
enhanced from the Kurokawa's human interaction models
(llun ueEnLO Aouap!£ dnplo
214
4900
Group Efficiency of Exploratory Group Formed from Leaders of Some Committee Meetings
(with condition for common consensus), GE ECW
for m = 0， GE = 0; for m = 1， GE = (n*p A (n-l)) * (p A n);
for m > 1, GE = (((n * p A (n-l)) * m) + (m * q A (m*(m-l)/2))) * ((p*q) A m) * (p A ((n-l)*m))
m = Network size of human group of exploratory leaders, n = Network size of every committee
meeting, q = Probability of the chemistry being good between a pair of leader members, p =
Probability of the chemistry being good between the chairperson & a member in a committee meeting
n
1
2
3
4
P
0.850
0.850
0.850
0.850
0.850
m \ q
0.800
0.800
0.800
0.800
0.800
0.000
0.000
0.000
0.000
0.000
1
0.850
1.228
1.331
1.282
1.158
2
1.665
1.670
1.433
1.136
0.859
3
1.426
1.281
0.953
0.649
0.419
4
1.079
0.876
0.566
0.331
0.182
0.805
0.583
0.326
0.163
0.077
6
0.614
0.388
0.186
0.079
0.032
0.475
0.258
0.105
0.038
0.013
8
0.366
0.170
0.059
0.018
0.005
9
0.280
0.110
0.033
0.009
0.002
10
0.211
0.071
0.018
0.004
0.001
4901
Group Efficiency of Exploratory Group Formed from Leaders of Some
Committee Meetings (with condition for common consensus)
— o— n
= 1
-z^-n
= 2
-e— n
= 3
= 4
2 3 4 5 6 7 8
Network size of human group of exploratory leaders
10
Figure 12.12 Group efficiency of an exploratory group formed from leaders of some
committee meetings (with condition for common consensus) as modified and
enhanced from the Kurokawa's human interaction models
(llun USEnLI) A0U3I0E 山 dnpls
215
5000
Success Probability of Exploratory Group Formed from Leaders of Some Committee
Meetings (with condition for common consensus), SP ECW
for m = 0， SP = 0; for m = 1, SP = p A n; for m > 1， SP = ((p*q) A m) * (p A ((n-l)*m))
m = Network size of human group of exploratory leaders, n = Network size of every committee
meeting, q = Probability of the chemistry being good between a pair of leader members, p =
Probability of the chemistry being good between the chairperson & a member in a committee meeting
n
1
2
3
4
5
P
0.850
0.850
0.850
0.850
0.850
m \ q
0.800
0.800
0.800
0.800
0.800
0.000
0.000
0.000
0.000
0.000
1
0.850
0.723
0.614
0.522
0.444
2
0.462
0.334
0.241
0.174
0.126
3
0.314
0.193
0.119
0.073
0.045
4
0.214
0.112
0.058
0.030
0.016
0.145
0.065
0.029
0.013
0.006
6
0.099
0.037
0.014
0.005
0.002
0.067
0.022
0.007
0.002
0.001
8
0.046
0.012
0.003
0.001
0.000
9
0.031
0.007
0.002
0.000
0.000
10
0.021
0.004
0.001
0.000
0.000
5001
0.900
0.100
Success Probability of Exploratory Group Formed from Leaders of
Some Committee Meetings (with condition for common consensus)
-e-n
= 1
-z^-n
= 2
-e— n
= 3
= 4
2 3 4
Network size of human group of exploratory leade
Figure 12.13 Success probability of an exploratory group formed from leaders of
some committee meetings (with condition for common consensus) ) as modified and
enhanced from the Kurokawa's human interaction models
>!IIICIBCIO.lcd sssons
216
In other words, if an organized crime group similar to committee meeting has
20 to 22 persons or more, then it starts to be inefficient. If the organized crime group
is similar to the exploratory group, then its inefficiency starts when the group has
five or more members.
Nevertheless, for the personnel in the CA, the situation is similar to the
committee meeting and getting 20 to 22 or more digital signatures from the CA
personnel is not that practical. For the introducer of trust of web, the situation is
similar to exploratory group. It is quite easy to get five of more digital signature to
certify a user's binding file. However, the trust level of introducer is limited to how
well the people know the introducer. It becomes quite impractical when other users
are asked if they know all the five or more introducers certifying a user's binding file.
Hence, other approach has to be implemented.
Up to here, the organized crime group, whether similar to committee meeting
and/or exploratory group, becomes inefficient when the number of group members is
more and hits a threshold. This is because criminals in an organized crime group are
normally lacking of a high level of trust among themselves. They normally try their
best to get rid of giving chances to other criminals to hold the evidence of their
criminal activities. The more members in an organized crime group, the harder it is
to be efficient. Furthermore, membership has to keep low to maintain a certain level
of profit sharing as reflected by the Sayan Chatterjee's article (No date) "Does
increased equity ownership lead to more strategically involved boards?".
A proof given to the Parkinson's Law is the time required to achieve a final
agreement on the works to be done tends to be more when more people are involved
and/or more time limit is given. This phenomenon is explained in articles Elliot
Aronson and Eugene Gerard (1966), "Beyond Parkinson's Law: The Effect of Excess
Time on Subsequent Performance"; Elliot Aronson and David Landy (1967)，
"Further Steps Beyond Parkinson's Law: A Replication and Extension of the Excess
Time Effect"; as well as David Landy, Kathleen McCue, and Elliot Aronson (1969)，
"Beyond Parkinson's Law: III. The Effect of Protractive and Contractive Distractions
on the Wasting of Time on Subsequent Tasks".
217
One more possible explanation is the longer time to achieve a common
agreement as in the BGP (Byzantine Generals Problem) together with the capability
to detect the faulty node. For organized crime group, all the members have to achieve
a common agreement and detect those possible faulty members before any action is
taken. As in the BGP for a very well-known fact, the larger is a network like the
human group, the more messages or time are needed to achieve the common
consensus. Therefore, to make the organized crime group to be inefficient, we have
to design a PKI similar to the Kurokawa's human interaction models.
Figure 12.11 illustrates the group efficiency of exploratory group formed
from leaders of some committee meetings without the condition for common
consensus among the members. This is an intermediate step to tell that when
common consensus among all the members is not needed, the group efficiency
increases as the members of exploratory groups and committee meetings increase.
Figure 12.12 illustrates the group efficiency of exploratory group formed
from leaders of some committee meetings with the condition for common consensus
among all the members. Here, all the personnel in the CA represent a committee
meeting, and each C A/introducer represents a member of the exploratory group.
Since other users only accept a MePKC digital certificate when all the
C A/introducer's digital signatures are verified, the organized crime group consisting
of the malicious CA and/or introducer has lower efficiency as the network size
increases. Figure 12.13 illustrates the success probability of exploratory group
formed from leaders of some committee meetings with the condition for common
consensus among all the members of the organized crime group.
12.3.3 Model to Boost up the Trust of MePKC Digital Certificate
It can be deduced that the more the criminals needed to succeed faking a
MePKC digital certificate, the lower is the success probability. One of the optimal
implementation is to have four (m = 4) or more groups of digital signatures for
binding file certification from the CA and/or introducers of trust of web, where each
CA contributes three (n = 3) or more digital signatures from its different personnel.
In this case, the success probability of the organized crime group is less than 6%.
218
5100
V
Method and system to boost up the trust level of MePKC digital certificate using
more than one certification authority (CA) and/or introducer of trust of web:
(1) First user creates an asymmetric key pair for MePKC digital certificate.
(2) First user binds the public key of the first user's asymmetric key pair, first
user identity, and other data, to create a binding file.
(3) First user sends the binding first to a first CA or introducer of trust of web for
certification to generate MePKC digital certificate.
(4) The first CA or introducer of trust of web authenticates the first user identity
using face-to-face checking of identity card or passport, or, if online transaction,
using the credit card number and bill.
(5) If first user identity is not authenticated, the first CA or introducer of trust of
web rejects the first user's certification application of MePKC digital certificate.
(6) Otherwise, if authenticated, the first CA or introducer of trust of web signs
and certifies the binding file as sent by the first user earlier by generating a first
digital signature later sent to the first user.
(7) The first's user MePKC digital certificate consists of the binding file and the
first digital signature from the first CA or introducer of trust of web.
(8) To increase the trust level of the first user's binding file, the user may send its
binding file again to a second CA or introducer for a second certification
application of a second MePKC digital certificate by repeating steps (3-6).
(9) The more the number of CA and/or introducer of trust of web certifying a first
user's binding file, the higher is the trust of the first user's binding file,
particularly, or MePKC digital certificate, generally.
(10) According to the Parkinson's Law, the coefficient of inefficiency is 20 to 22
persons for a human group meeting together to achieve a target.
(11) According to the derivation of Parkinson's Law, the trust level of this
method reaches a critically safe level when the number of members of an
organized crime is more than 20 to 22.
(12) When the Kurokawa's human interaction model is simulated for the
organized crime to create fake MePKC digital certificate, one of the optimal
implementation is to have four or more groups of digital signatures for binding
file certification from the CA and/or introducers of trust of web, where each CA
contributes three or more digital signatures from its different personnel.
5101
V
Other users like a second user verifying the first user's MePKC digital certificate:
(1) A second user receives the first user's MePKC digital certificate(s) consisting
of one binding file and digital signature(s) of the CA and/or introducer(s) of web
of trust.
(2) If all the digital signature(s) are verified, second user accepts the first user's
MePKC digital certificate.
Figure 12.14 Method and system to boost up the trust level of MePKC digital
certificate by using more than one certification authority (CA) and/or introducer of
trust of web
219
Figure 12.14 illustrates the operations of the method and system to boost up
the trust level of the MePKC digital certificate. Now, the first PKI risk informed by
Carl Ellison and Bruce Schneier (2000) on "Who do we trust, and for what?"
questioning on how well the CA maintains its private keys well and the third risk on
"How secure is the verifying computer?" questioning on the possibility of attacker
adding its own public key to the list of certificate verification, can also be improved
by having more than one C A/introducer certifying a digital certificate. This is
possible because users can generate their own asymmetric key pairs. The CA or
introducer of trust of web may be a government authority, and people working in the
fields of religion, law, police, security, politics, army, finance, diplomacy, etc., who
have a high trust level in the society like judge, Commissioner for Oaths, lawyer, etc.
220
CHAPTER 13 MePKC TIMESTAMPING SCHEME FOR
EVIDENCE OF INTELLECTUAL PROPERTY
(IP) ORIGINALITY
13.1 Proof of Copyright Ownership Using Digital Timestamp in Malaysia
Copyright is an intellectual property needing no registration generally for an
original work of authorship like literary, artistic, musical, or computer program
affixed in any tangible medium of expression. It lasts for an author's life plus 50
years in Malaysia. Without registration, the author has to prove its time of creation
using other methods. In Malaysia, proof of copyright and priority is generally
established by way of a statutory declaration (SD). The cost is low and the
processing fast. However, its disadvantages are low confidentiality and possible
collusion between the author and Commissioner for Oaths. Here, digital
timestamping under Malaysia's Digital Signature Act and Regulations (Act 562) is
proposed to prove copyright ownership in electronic works (Lee, Radhakrishna &
Khaw, 2007). Digital timestamping providing lower cost, faster processing, higher
confidentiality, better security, and management, is a variant of digital signature.
13.1.1 Introduction
Intellectual properties (IPs) have become increasingly important in developed
countries. Taplin (2004) reported that IP assets accounted for up to 70% and 40% of
market values of all corporate assets in United States and Japan, respectively.
Chandran (2007) stated that for any company, and especially for pharmaceutical
companies, IP is more valuable than any of its tangible physical assets, where IP
constituted more than 80% of the total revenues of any company while tangible
assets accounted for only 20%. Again, Ocean Tomo LLC (Ocean Tomo, No date;
Wikipedia Contributors, 2008d) figured out that the components of S&P 500 market
value in 2005 also had 79.7% to be intangible assets.
Among the IP, patent has the largest economic value. In fact, IP is a group of
rights consisting of copyright, patents, registered designs, trademarks, and know-how
221
(Curzon, 1998; Multimedia Development Corporation Sdn Bhd (MDC), 2002a,
2002c, 2002d, 2002e).
Certain IPs require official registration while others don't. Copyright requires
no registration in Malaysia. This interdisciplinary paper looks at various methods for
proof of copyright, and seeks to promote the idea of digital timestamping as a viable
solution to evidentiary requirements. Both the legal and technical aspects of digital
timestamping are discussed.
In the US, in order to claim statutory damages, a copyright has to be
registered at the US Copyright Office. The statutory damages are pre-established
damages for cases where a correct sum is deemed difficult to be calculated
(Wikipedia Contributors, 2006). The basic level of statutory damages in US ranges
from USD$750 to USD$30,000 per work at the discretion of court (World
Intellectual Property Organization (WIPO), No date; Wikipedia Contributors, 2007e).
In Malaysia, a statutory declaration (SD) or affidavit is commonly used to
prove the copyright authorship under the Copyright Act 1987 (Act 332) (MDC,
2002a). Copyright offender in Malaysia is subject to a fine not exceeding
MYR$25,000 or to imprisonment not exceeding 3 years or to both under Copyright
Act 1987 Section 43.
Digital timestamping is a cryptographic scheme being a variant of digital
signature. In Malaysia, it is enforced under the Digital Signature Act 1997 and
Regulations 1998 (Act 562) (MDC, 2002b). Currently, Malaysia has only 1
registered service provider of digital timestamping, i.e. MSC Trustgate.com Sdn.
Bhd. (47823 1-X), a subsidiary of Multimedia Development Corporation Sdn. Bhd
(MDeC, which is previously known as MDC) (Malaysian Communications and
Multimedia Commission (MCMC), No date).
Haber and Stornetta (1990， 1991) firstly proposed a digital timestamping
scheme with its described applications for scientific priority of patent inventions [9-
10]. Here, we wish to extend its application to proof of copyright authorship and/or
ownership in the Malaysian perspective. Cost estimation and comparisons with other
methods are presented.
222
13.1.2 Related Works: Copyright
In Malaysia copyright is governed by the Copyright Act 1987， which extends
protection for a qualified person to all types of literary, musical, artistic works, film,
sound recording, broadcasts, derivative work, published edition, and live
performances (Khaw, 2001).
Copyright subsists in original literary (including computer programs),
musical, or artistic works, sound recordings, films, broadcasts or cable programmes,
architectural plans, and the typographical arrangements of published editions of
works (Curzon, 1998).
In Malaysia, the copyright owner is given the right to reproduce, prepare
derivative works, distribute, communicate their works, perform, or rent their works
for the duration of the author's lifetime plus 50 years after his death. In European
Union, the protection period is 70 years after the death of the last author; whereas
corporate authorship is 70 years from the year the work is created. In the US, the
Sonny Bono Copyright Term Extension Act of 1998 extends the protection period of
individual works to 70 years after the death of the last author; whereas corporate
authorship is 120 years from the date of creation or 95 years after publication,
whichever is shortest. Hence, the security design of timestamping scheme has to last
for about 100 years in Malaysia.
It is transmissible by assignment or will as personal property. Copyright
exists automatically once it is expressed in any tangible material form, e.g. if one
doodles on a piece of paper while chatting on the phone, then that doodle is
automatically protected without the need for any registration. However, if one were
to compose a tune in one's head and hum it in the shower that is not protected as it
has not been reduced to material form. Were it to be recorded then it would be
protected.
The first owner of a copyright work is the author. Then, copyright may be
assigned or transmitted. In cases where a work was created under a contract of
employment or pursuant to a commission, the copyright in the work is transferred to
the employer or commissioner, as the case may be. It must be noted that copyright is
vested in the copyright owner and not the author, unless he is also the owner.
223
There is no registration system for copyright in Malaysia (Intellectual
Property Corporation of Malaysia (Perbadanan Harta Intelek Malaysia) (MylPO), No
date). In the event of any copyright dispute, authorship and/or ownership would have
to be proved by the party initiating the action. Internationally, the Berne Convention
for the Protection of Literary and Artistic Works (WIPO, 1979; Wikipedia
Contributors, 2007i) in 1887 set out the scope of copyright protection, applicable to
all member countries. It was formed in order to formulate greater uniformity in
copyright law among countries and to give copyright owners certain minimum levels
of copyright protection without the requirements of registration in member countries
(Articles 1 and 2 (WIPO, 1979)). The Berne Convention requires a signatory country
to recognize a copyrighted work of authors from other signatory countries in the
same way as it recognizes the copyright of its own national or domiciliary.
13.1.3 Related Works: Methods of Proving Copyright Ownership
Currently, there are six methods to prove the copyright ownership as follows:
(i) Self-addressed envelope with original work mailed back to oneself;
(ii) Log book;
(iii) Lodging a copy with solicitor or bank;
(iv) Hashing;
(v) Registration and/or preregistration at copyright office; and
(vi) Statutory declaration (SD).
For self-addressed envelope with original work mailed back to oneself, this is
perhaps the oldest method being used to prove the scientific priority and copyright
authorship or ownership. The original work is inserted into a self-addressed envelope,
which is then sealed and mailed back to the author. The postal time stamp acts as a
proof for the creation time. Whenever there is a case of copyright infringement, the
sealed envelope is opened in public to settle the dispute. However this is known as
the "poor man's copyright" as there is scope for abuse. There is nothing to prevent
224
tampering or pre-po sting envelopes to oneself and inserting allegedly original works
into the envelope later. As such they will not stand up as credible evidence in court
(UK Copyright Service (UKCS), No date a).
For log book, an author may keep a habit like an inventor and scientist, where
a log book is properly recorded and endorsed by one or more third parties regularly.
In case of dispute, the log book and the endorsers can provide the proof of authorship.
For lodging a copy with solicitor or bank, the party, who endorses a log book,
can be specifically a solicitor as in log book. Sometimes, a copy of the copyrighted
works can be lodged with a solicitor and/or bank, who act as copyright witnesses.
However, copyright witnessing is not their main concern and they are unlikely to
understand the essential implications of the service (UKCS, No date a). Other
weaknesses are data loss due to corruption and disaster as well as loss of evidence for
future cases.
For hashing, hashes are cryptographic operations performed to transform one
or more fields into a unique fixed bit stream as the contents of a piece of digital
evidence (Oppliger & Rytz, 2003; Maurer, 2004) can be preserved. These
mechanisms ensure that digital information "has not been altered." Cryptographic
hashes can be used to protect evidence from tampering for long periods of time.
Using algorithms such as SHA-2 and RIPEMD-256, content integrity can be attested
to for many years even under distributed brute force attack. Thus, by applying
cryptographic hashes and digital signature technology to the problem of evidence
preservation, the "who" and "what" of digital data can be identified and maintained.
However this still leaves time to be established to claim scientific priority and
copyright evidence (Duren & Hosmer, 2002).
For registration and/or preregistration at copyright office, in UK and US,
there are copyright offices managing the copyright registration as strong evidence
upon a fee payment. In UK, copyright registration service is offered by The UK
Copyright Service (UKCS) (UKCS, No date b). For online registration, UKCS
charges GBP$35 for 5 years or GBP$60 for 10 years per work, where it covers an
upload up to 10MB per registered work and upload over 10MB are subject to an
additional fee of 2 pence (= GBP$0.02) per additional 1MB or part thereof up to
225
virtually unlimited size. For postal registration, UKCS charges GBP$40 for 5 years
or GBP$70 for 10 years per work, plus different additional fees according to the
formats of hard copies, up to a maximum size of 8.5 GB. Meanwhile, the US
Copyright Office (USCO) offers registration and preregistration services.
Preregistration is for unpublished work, where creation of work has begun aiming for
commercial distribution, e.g. motion picture, musical work, sound recording,
computer program, book, or advertising photograph. Preregistration is not a
registration and its fee is USD$100. After a work is published, registration at USCO
is required and the electronic filing of a basic copyright registration is USD$35 and
paper application is USD$45 (USCO, No date).
For statutory declaration (SD)， this is the most commonly used method of
establishing copyright in Malaysia. By Malaysia Copyright Act 1987 Section 42， an
affidavit or statutory declaration made before a Commissioner for Oaths (Curzon,
1998) or Notary Public and asserting that copyright subsists in a particular work and
the person named therein is the copyright owner may be admitted as prima facie
proof of the facts stated therein. This is a convenient and economical way of proving
copyright and ownership. Section 42 places the burden on the infringer to dispute
and challenge the prima facie evidence adduced by the copyright owner
(International Intellectual Property Alliance (IIP A), 200 1 ). Section 42 was
introduced to facilitate proof of copyright and ownership in cases involving foreign
copyright owners.
There are some problems for SD like unreliability of information in the SD
itself, need to verify information in the SD itself, challenges to information in the SD,
and SD proof for patent invention is hard to establish. In the US, the legal case of
laser inventor Gordon Gould (Taylor, 2002; Wikipedia Contributors, 2007n) uses SD
and it has lasted for about 30 years. Hence, it is inefficient for IP evidence, especially
for patent.
However in practice, there have problems with the admission of Section 42
affidavits. Courts are cautious with respect to presumptions, and have required
prosecutors to prove subsistence issues through other documents such as record
company receipts of first publication, letters of authority, or sometimes even live
226
testimony of right holder representatives (CLJ Legal Network Sdn. Bhd.， 1998) (CLJ
- Crime, Law, and Justice). Failure to comply with these requirements has in some
cases led to acquittals or the rejection of the affidavits as evidence of ownership. The
root of the problem has been that not many judges are well versed in Copyright law
and defence counsels are quick to challenge any actions on technical grounds.
Further weaknesses are its low confidentiality and possible collusion between the
author/owner and Commissioner for Oaths.
13.1.4 Digital Timestamping Scheme to Prove Copyright Ownership
Ismail (200 1 ) discussed some timestamping schemes for Malaysian
applications. Among them, linking scheme (Haber & Stornetta, 1990， 1991) and tree
scheme (Benaloh & de Mare, 1991) limit the power of timestamping authority (TSA).
Here, a modified scheme is proposed to limit the TSA power but has simpler
implementation as in Figures 13. 1-13. J. The TSA cannot collude with a user to
backdate a time stamp more than one day using publication of dany superhash at a
repository authority like newspaper.
13.1.5 Malaysia Digital Signature Act 1997 and Regulations 1998 (Act 562)
A "digital signature" is a transformation of a message using an asymmetric
crypto system such that a person having the initial message and the signer's public
key can accurately determine (Section 2 (MDC, 2002b)):
(i) whether the transformation was created using the private key that
corresponds to the signer's public key; and
(ii) whether the message has been altered since the transformation was
made.
Thus it is able to identify the signer and also pick up evidence of any
tampering. Section 64 deems a digitally signed document to be a written document as
if it had been written on paper if:
227
(i) it bears in its entirety a digital signature; and
(ii) that digital signature is verified by the public key listed in a certificate
which:
(1) was issued by a licensed certification authority; and
(2) was valid at the time the digital signature was created.
0.0 Initialization vector, IV = a random bitstream for the first day or yesterday
superhash for second day and beyond. Let i = 1， 2， 3， ...， N， where N is the
number of total time stamp users for one day, and j = 1， 2， 3， ...， oo.
1.0 For j-th day with N users.
1.1 User Ui creates hash Hn of a digital document D.
1.2 User sends Hn to timestamping authority (TSA).
1.3 TSA creates hash from the concatenation Cn of Hn and a time stamp Ti
having the present time synchronized to an Internet time server like
"time.windows.com" or "time.nist.gov".
1.4 TSA signs the H 2 i using its private key K pte and creates hash H 3i as signed
timestamp.
1.5 TSA stores the concatenation of H 3i and IV
1.6 TSA sends Cii to Ui and repository authority R.
1.7 User Ui verifies the signed timestamp H 3i using Ti and TSA's public key K pub
as in Fig. 2.
1.8 If H3i is verified, user Ui accepts the H^; otherwise, Ui rejects the H3i and goes
to step 1.2.
2.0 TSA creates a daily pre- superhash PHj from the concatenation of ordered
on j-th day.
PHj = hash(C 2 i II C 22 II C 23 ᅵᅵ… ᅵᅵ C 2i I ᅵ… II C 2N )
3.0 TSA creates daily superhash SHj from the concatenation of IV and PHj on j-th
day.
3.1 If j = 1, IV = a random bitstream; otherwise if j > 2， IV = SHj— 1 .
3.2 SHj = hash(IVIIPHj)
3.3 TSA stores the superhash SHj.
4.0 TSA sends the SHj to repository authority R for deposit and publication to
limit the TSA power.
5.0 Go to step 1.0 for a following day.
Figure 13.1 Generation of timestamp and superhash
228
0.0 Digital or electronic document D, concatenation of signed time stamp
and time stamp IV
1.0 User creates hash H v n of a digital document.
2.0 User creates hash H V 2i from the concatenation C v h of H v n and IV
3.0 User applies TSA's public key K pub on signed timestamp H 3i to create
decrypted signature H V 3i.
4.0 If H V 2i = H V 3i, the signed timestamp H 3i is verified and the user accepts it;
otherwise if H V 2i ^ H V 3i, signed timestamp is rejected.
Figure 13.2 Verification of a signed timestamp
0.0 First day IVi or yesterday superhash SHj_i, daily superhash SHj,
concatenations C2i of signed time stamps H 3i and time stamps Ti on j-th day
from repository R. To-be-verified concatenation C V 2i supplied by a user Ui.
1.0 User Uk checks the presence of C V 2i among the Cii in the R's record.
2.0 If C V 2i equals to one of the C^, then go to step 3.0; otherwise, user Uk rejects
the C V 2i and ends.
3.0 User Uk creates a daily pre- superhash PHj from the concatenation of ordered
C2i on j-th day.
PHj = hash(C 2 i II C 22 II C 23 II ... II C v2i II ... II C 2N )
4.0 User Uk creates a to-be-verified daily superhash SH V j from the concatenation
of IV and PHj.
4.1 If j = 1， IV = IVi; otherwise if j > 2， IV = SHj 小
4.2 SH V j = hash(IV II PHj)
5.0 If SH V j = SHj, then user Uk verifies the C V 2i; otherwise, user Uk rejects the C V 2i.
Figure 13.3 Verification of daily superhash to limit TSA power
Further under Section bo, a certificate issued by a licensed certification shall
be an acknowledgement of a digital signature verified by reference to the public key
listed in the certificate, regardless of whether words of an express acknowledgement
appear with the digital signature and regardless of whether the signer physically
appeared before the licensed certification authority when the digital signature was
created, if that digital signature is:
(i) verifiable by that certificate; and
229
(ii) affixed when that certificate was valid.
Section 67 presumes the validity and truth of the contents of the certificate
issued by a recognized Certification Authority or Repository as to the digital
signature being that of the subscriber. Apart from digital signatures, Section 70 also
recognizes digital timestamps. A "time-stamp" means (Section 2 (MDC, 2002b)):
(i) to append or attach to a message, digital signature or certificate a
digitally signed notation indicating at least the date, time and identity
of the person appending or attaching the notation.
Further Section 62 of the DSA Regulations provides that a recognised
date/time stamp service shall:
(i) on receipt of a document for time- stamping, immediately time- stamp
the date and time of its receipt on the document and digitally sign the
time- stamp; and
(ii) at the end of each business day cause to be published only the hash
result of the document in at least one recognised repository all
documents time- stamped by it in that day.
Thus a digital timestamp is able to provide evidence of "who", "what", and
"when" accurately to establish priority of copyright. It helps the author to prove
when the idea was expressed - this is the moment when copyright protection begins
and is especially effective for protecting prepublication work. The third-party
timestamp and certificate provides evidence to help resolve or avoid legal disputes
without the need for costly and time-consuming litigation.
13.1.6 Malaysia Evidence Act 1950 (Act 56)
Amendments to the Evidence Act 1950 (Act 56) allowed the admissibility of
"computer-generated documents" (International Law Book Services (ILBS), 2003).
230
Under Section 3， a matter recorded, stored, processed, retrieved, or produced by a
computer is a "document". Section 62 explains that a document produced by a
computer is primary evidence. Section 78A accepts public documents produced by
computers. Section 90A admits in evidence documents produced by computers
subject to certain conditions. Timestamp produced by computers is a document that
can act as primary and prima facie evidence.
13.1.7 Security Analysis
The security of the proposed digital timestamping scheme depends on the
lowest security strength among the hash algorithms and digital signature scheme on
timestamp. Hash algorithms like SHA-256, SHA-384, and SHA-512, have no key
and the proposed bits of security range from 128 to 2^0 bits with protection periods
at 30 years to foreseeable future, respectively, based on the symmetric key scale.
For digital signature scheme of timestamp, it needs a key pair to sign and
verify a digital signature. Due to human memory limits, memorizable private key is
currently not practical until the invented big secret creation methods are proposed in
this research project. Hence, encrypted private key or two-factor authentication using
a symmetric key and a token is normally used. In other words, the weakest point of
digital timestamping scheme is at the digital signature protocol, which depends on
the safety of symmetric key encrypting a private key and the availability of token.
The security of the proposed timestamping scheme in Figures 13.1-13.3 is
similar to the linking (Haber & Stornetta, 1990, 1991) and tree schemes (Benaloh &
de Mare, 1991) because only the superhash generation is different. The daily
superhash in the proposed scheme here is generated from the concatenation of
yesterday superhash and the concatenation of all the timestamps at the present day.
These schemes rely on the security of hash function and digital signature for
the individual timestamp. For the security against the malicious TSA, it relies on the
aaily superhash safely kept at an independent repository authority. As long as there is
no collusion between the TSA and repository, malicious TSA can be detected. If
there exists collusion between TSA and repository, public can detect this event by
keeping a published daily superhash for future verification. Another alternative is to
231
increase the number of repository. Table 13.1 compares the proposed scheme with
the linking and tree schemes.
Table 13.1 Comparisons of our scheme with linking and tree timestamping schemes
Feature
Linking
Scheme
Tree Scheme
Proposed
Scheme
Signature Scheme
Same. Equal security strength.
Single Timestamp
Same. Equal security strength.
Number of Computer Units
One
Many
One
Generation Speed of Superhash
Slow
Fast
Moderate
Malicious TSA
Same. Equal security strength.
13.1.8 Computation Load
The computation load is heavy during the signing process and light during the
hashing process. There is only one signing process for one timestamp. Meanwhile,
there are many hashing processes for timestamp and superhash. Hence, the
computation load is efficient and similar to the linking scheme and tree scheme
(Ismail, 2001). The signing process is also fast enough because it is signing a hash
value rather than a whole digital file.
13.1.9 Implementation Costs
The implementation cost of this digital timestamping scheme is estimated
based on the one-time and annual operating costs as in Table 13.2.
The USCO receives 600,000 copyrighted works annually from a 300 million
population. Malaysia with a population of 27 million is estimated to have 54,000
copyrighted works annually by assuming the US and Malaysia have same level of
creativity. Let one timestamp to be charged at MYR$5 for 5-year repository period,
then a total of MYR$270,000 can be obtained from the sale of timestamps. One can
breakeven within the first year. To absorb the hidden cost and inflation as well as to
maintain a fixed price per timestamp for about 5 to 10 years, a fee of MYR$10 per
timestamp shall be charged for a repository period of 30 years.
232
Table 13.2 Cost estimation table of a digital timestamping scheme in Malaysia on 8
June 07
Item
T T ' j.
Unit
Cost (RM)
Repository Licence
- establishment stage
2,500
- operation stage
1
2,500
- granting tee
つ f\ r\f\r\
30,000
allllUd.1 U Uv-/A a Llllg ICC
7 SOD
ᄂ /IglLdl L1111C o LalllU 니ᄂ Cll ᄂ C
- establishment stage
2,500
- operation stage
1
2,500
- granting fee
30,000
- annual operating fee
2,500
Computer
4
10,000
Windows Vista Business Edition
4
5,080
Norton Antivirus 2007 (1 year)
4
558
ZoneAlarm Internet Security Suite (3 PCs) (1 year)
2
349
Spybot - Search 7 Destroy 1.4
4
2,000
InstallShield 12 Express (Windows) + Visual Studio .NET
Professional 2005
9
1 ,001
Corporate Web Hosting (1 year)
1
1,000
Furniture
1
6,000
Annual Company Secretary Fee
1
3,000
Annual Accounting Fee
5,000
Annual Auditing Fee
5,000
Annual Office Rental
24,000
Annual Advertisement Fee
12,000
Annual Transportation Fee
3,000
Annual Legal Fee
6,000
Annual Engineer Salary
2
72,000
Annual CreditCard Commission (VISA & MasterCard @
2.35%)
54k
1,269
Total (one-time cost + annual fee, where n is year)
100,741 +
138,176n
N.B. USD$1.00 = MYR$3.4850 (RM) (The Star Online, 2007)
233
A regular renewal can be made. Otherwise, longer protection periods need
higher fees such as MYR$20 per timestamp for 100 years subject to the TSA
sustainability.
13.1.10 Comparisons
Table 13.3 compares the proposed method with other methods to prove
copyright authorship/ownership. In Malaysia, an SD is charged at MYR$4 for the
first copy and its following copies. The duplication-free and travelling-free factors
make timestamp cheaper. Other advantages of copyright evidence using digital
timestamp are independent evidence, contractual contingency, confidentiality, fast
processing, limited authority power, low cost, and better management.
Table 13.3 Comparisons of methods to prove copyright authorship/ownership
Features
Self-Addressed
Posting
Log Book
Lodging a Copy
with Bank, etc.
Hashing
Registration
Statutory
Declaration
Digital
Timestamping
Independent Evidence
N
N
Y
Y
Y
Y
Y
Contractual
Contingency
N
N
N
N
Y
N
Y
Confidentiality
N
N
N
Y
Y
N
Y
Duplication-Free Cost
N
N
N
Y
N
N
Y
Travelling-Free Cost
N
N
N
Y
Y
N
Y
Fast Processing
N
N
N
Y
N
Y
Y
Limited Authority
Power
N
N
N
N
Y
N
Y
Low Cost
Y
Y
N
Y
N
Y
Y
Better Management
N
N
N
Y
Y
N
Y
Low Complexity
Y
Y
Y
N
N
Y
N
N.B.: N = No, Y = Yes.
234
13.1.11 Conclusion
From both the legal and technical perspectives, digital timestamping offers
comparatively better proof, security, efficiency, and economy than the other methods
discussed, especially SD. This section mainly proposes a new alternative to prove
copyright ownership from the legal aspect using a cryptographic timestamping
scheme. The proposed timestamping scheme is slightly modified from the previous
schemes.
13.2 IP Evidence Using Digital Timestamp
13.2.1 Introduction
Among the IPs, patent and copyright are the sources of income. Gregory
Aharonian (No date) listed out the incomes due to patent and copyright in the
website of STO (Source Translation and Optimization). Gordon Gould (Wikipedia
Contributors, 2007n)， who is the inventor of laser, has a thirty- year patent war of
lawsuit to claim his invention for patent (Taylor, 2002). The evidence to prove the
creation time is critical for his patent filing because the US follows the first-to-invent
system. Sarcastically, Gould is not awarded the Nobel Prize in Physics for his
invention of laser but there are many other Nobel laureates based on laser.
In a book excerpt by Nick Taylor (2002):
"In 1957 Gordon Gould, then an obscure physicist and perennial graduate
student, conceived one of the revolutionary inventions of the twentieth century ― the
laser. But before he could submit a patent application, a prominent professor of
physics whose office was next door to uould's filed his own laser patent claims.
Gould fought to reclaim the rights to his work, beginning a battle that would last
nearly thirty years."
Here, the notary system has weakness. Otherwise, the lawsuit would not have
been prolonged to 30 years. The disadvantages are low confidentiality and possible
collusion between the inventor/author and Commissioner for Oaths. In the digital era,
it is also hard to prove the creation time of a copyrighted work. Neea is required to
235
differentiate between originality and plagiarism. Consequently, digital timestamp
service is proposed here to provide IP evidence in Malaysia and other geo-political
regions, particularly patent and copyright.
The suitability of timestamp evidence for patent filing system is investigated.
There are two patent filing systems: First-to-invent and first-to-file. These patent
filing systems have the mixed advantages and disadvantages like first inventor in the
interference case, possible idea accumulation towards commercial value, reasonable
time for prototype development, notary service, patent translation costs, patent filing
costs, and intentional delay of patent filing.
Based on the good and bad points, a hybrid system may be introduced, which
is called "limited first-to-invent patent filing system", where the period between the
first date of invention claim and first filing date is within a reasonable figure like one
to two years. The legislation of the related laws in the fields of cyberlaw and
intellectual property (IP) law is needed to enforce the IP evidence using digital
timestamp as proposed here.
13.2.2 Related Works: Patent Filing Systems
In this world, there are two patent filing systems: First to invent, and first to
file. By the last decade, there are only three countries in the world implementing the
first-to-invent patent filing system (Inventors Assistance League, No date; Coster,
2002; Wikipedia Contributors, 2007j); whereas the other countries are all
implementers of first-to-file patent filing system. Till today, the USA is the only
country implementing the first-to-invent patent filing system.
Nevertheless, now even the USA is on the brink of patent laws reformation
since the last decade (Samuelson, 2004; Mossinghoff, 2005; Wikipedia Contributors,
2008b, 2008e) for the issues like patent filing system, number of core and edge
invention claims in a patent, unity of invention (Wikipedia Contributors, 2008ba)，
and inventorship (Wikipedia Contributors, 2008ao).
First to invent means the first inventor has the right to file for patent first
upon proving the first date of invention claim whenever there is an inventor
236
interference case. First to file means the inventor, who firstly files for the patent
application throughout the world, has the right to claim as the inventor and the filing
date is the invention priority date.
13.2.3 Related Works: Digital Timestamping Scheme
The digital timestamping scheme proposed in Section 13.1 to prove copyright
authorship/ownership can be extended to other IPs like patent here. Please refer to
Sections 13.1.9 and 13.1.10 for the implementation costs, as well as comparisons for
advantages and disadvantages, respectively.
For clarity, the added advantages includes confidentiality, infinite
complimentary duplication of timestamp, savings on transportation and operating
costs, lower overall cost, faster processing, better management, and limited authority
power of timestamping service provider. The timestamping authority (TSA) has
limited power that it cannot backdate because of the daily superhash stored regularly
in a trusted repository everyday. In case the TSA is malicious, the accuracy of the
timestamp can be trusted at the unit of day. To increase the fault tolerance to
malicious repository, the number of repositories can be increased and resorts to the
Byzantine Agreement Protocol (BAP) of the Byzantine Generals Problem (BGP).
13.2.4 Proposing Limited First-to-invent Patent Filing System
IP evidence with digital timestamp is a timely proposal in view of the
inadequacies of existing methods to prove IP creation time. Additionally, the advent
of computing technology has paved the way for more reliable and efficient means to
digitally protect information (IP inclusive).
At the same time, the proposed service will prevent patent lawsuits similar to
Gould's from recurring, which is based on first-to-invent system in the USA rather
than first-to-file system in the rest of the world.
Considering the advantages and disadvantages of these systems (Inventors
Assistance League, No date; Coster, 2002; Mossinghoff, 2005)， like first inventor in
the interference case, possible idea accumulation towards commercial value,
237
reasonable time for prototype development, notary service, patent translation costs,
patent filing costs, and intentional delay of patent filing, a hybrid system shall be
introduced, which is a limited first-to-invent patent filing system, where the period
between the first date of invention claim and first filing date is within a reasonable
figure like one to two years.
In a book excerpt by Burgess and Power (2008):
"The U.S. Chamber of Commerce estimates that counterfeit and pirated
products account for 5 percent to 7 percent of the global economy, and results in the
loss of more than 750,000 jobs and approximately $250 billion in sales to the United
States alone
The threats of economic espionage and intellectual property (IP) theft are
global, stealthy, insidious, and increasingly common. According to the U.S.
Commerce Department, IP theft is estimated to top $250 billion annually and also
costs the United States approximately 750,000 jobs. The International Chamber of
Commerce puts the global fiscal loss at more than $600 billion a year."
Limited first-to-invent patent filing system has all the features of first-to-
invent patent filing system except that the first date of invention claim and first filing
date is limited to be within a reasonable figure like one to two years, depending on
the further research studies for better optimization. The main aim of this limited first-
to-invent system is for data protection (Ha skin, 2008) to resist the prevalent hacking
activities (McClure, Scambray & Kurtz, 2001; Beaver, 2004; Pankaj, 2005; Berinato,
2007a, 2007b, 2007c, 2007d， 2007e, 2007f; Wikipedia Contributors, 2007q, 2008n;
McMillan, 2008b) and economic espionage (Fialka, 1999; Fink, 2002， 2003;
McNamara, 2003; Nasheri, 2004; Burgess & Power, 2008). Hacking and economic
espionage are the main killing weaknesses of the first-to-file patent filing system.
Beaver (2004) reported that a networked computer without proper firewall
(Ogletree, 2000) settings would be hacked within 30 minutes. Yet in the latest news,
Markoff (2008) informed that the hacking period dropped to less than 5 minutes after
a hacker had operated for 30 seconds to access a prey computer. This reflects how
serious and dangerous the current computer communications network security
( Stalling s, 2000) is in this networked info-computer era.
238
Identity theft can happen when a hacker copies a prey's computer data as disk
image (Wikipedia Contributors, 200 8bf) using disk cloning software (Wikipedia
Contributors, 200 8be), and then put the disk image into a second computer and
modify, add, delete, etc. on some contents, which creates a second type of zombie
computer (Wikipedia Contributors, 200 8bj) and/or botnet (Wikipedia Contributors,
2008bd， 2008bi). This second type of zombie computer, when connected to the
Internet, can fool other prey hackers watching this zombie computer version 2. Of
course, if there are any confidential information, business secret, and other intangible
assets, in the prey computers, then they shall be considered as disclosed and released
to the public domain.
Tables 13.4-13.6 show the advantages and disadvantages of first-to-file, first-
to-invent, and limited first-to-invent patent filing systems, respectively. From these
tables, it can be concluded that limited first-to-invent patent filing system shall be
highly appreciated and MePKC digital time stamp shall be used as the IP evidence.
Table 13.4 Advantages and disadvantages of first-to-file patent filing system
Advantages
[1] No interference case to determine the first author
[2] Eliminate intentional delay of patent filing
Disadvantages
[1] No chance for idea accumulation towards commercial value
[2] Less number of claims per patent application
[3] Very high patent filing cost
[4] Risk for hacker stealing the electronic file that can be easily
copied in this networked computer world
Table 13.5 Advantages and disadvantages of first-to-invent patent filing system
Advantages
[1] Possible for idea accumulation towards commercial value
[2] Has time for prototype development
[3] More claims per patent application
[4] Lower patent filing cost
Disadvantages
[1] Public notary service used to prove the first inventor is not
efficient and practical by referring to the exemplary case of the
laser inventor of Gordon Gould which has lasted for 30 years
[2] Interference case to determine the first inventor is hard
[3] Intentional delay of patent filing for public disclosure
239
Table 13.6 Advantages & disadvantages of limited first-to-invent patent filing system
Advantages
[1] Possible for idea accumulation towards commercial value
[2] Has time for prototype development
[3] More claims per patent application
[4] Lower patent filing cost
[5] Digital timestamp is more efficient and practical than public
notary service.
[6] Interference case to determine the first inventor is easier
[7] Reasonable delay of patent filing for public disclosure
Disadvantages
[1] Need survey and research to determine the window period
between the first date of invention claim and first patent filing
date, which is suggested to be from 1 to 2 years.
[2] Need international revision of patent law, especially PCT under
the WIPO that may take a long time
240
CHAPTER 14 HACK-PROOF DATA STORAGE USING
INNOVATED DIP SWITCH
14.1 Abstract
A dual in-line package (DIL/DIP) switch has been modified to collectively
link all the poles using a single actuator and called secure DIP switch. The actuator
can be a raised/recessed slide, raised/recessed rocker or piano-type (aka side/level),
selectively switching on or off one/two groups of poles oppositely. A specific
inventive application is when a 10/ 12- way secure DIP switch is integrated with two
modular jack RJ45 sockets and a second storage device preferably via USB
connection, a secure data storage resisting the computer hacking in a malicious
computer network is created. This new component is simple, cost-effective, and
hack-proof. Yet a novel variant is N1PST+N2PST DIP switch with reverse activation.
14.2 Introduction
Hacking or cracking into a computer from a malicious computer network
(Wikipedia Contributors, 2007q， 2008n) is a great threat to the information security
of private and confidential data in this electronic society. History of hacking and
cracking can be traced (Wikipedia Contributors, 2008n). To resist the hacking and
cracking, network settings and firewall software (Ogletree, 2000) are among the
available best tools. However, these tools are complicated and not user-friendly to a
networking novice like common Internet user. They are only good to network
administrator who has undergone training and/or understood the operating manual.
In other words, network settings and firewall software are excellent at the
server side but not the client side. Technical difficulty and affordable cost are two
main factors discouraging the users to adopt these two anti- hacking approaches
effectively. Furthermore, end users normally do not require data sharing via web
hosting like server. This indicates that private and confidential data of end users can
actually be partitioned from the data without security concern. For more information
on the imperative demand of hack-proof data storage, please refer to Section 13.2.4.
241
In addition to the financial loss of confidential information and business
secret, there are cronies of organized crime using the hacked secrets, flash mob
approach, and sound snatching to conspire for more serious crimes like to worsen a
good human relationship and/or to fasten a cheating human interaction. Married
couples may be made divorced. Lovers may be made suspicious between themselves.
Relatives, friends, colleagues, and organization members may be made trust-less and
negatively emotional. Cheaters may succeed to establish trust, cultivate positively
false emotion, and build a dishonest relationship leading to a marriage for sharing or
even controlling the power, wealth, reputation, and fame of a single man or woman
with good social status. In short, the criminals may cheat for sex, trust, emotion,
power, money, and assets.
Here, method and device are proposed to secure a hacking-free (or hack-
proof) data storage for end users. This method uses a new component called secure
DIP switch integrated with two modular jack sockets and a second storage device
like hard disk drive (HDD) or USB (Universal Serial Bus) flash drive. Private and
confidential data is stored in the second storage device. Secure DIP switch controls
the normal networking mode while it is switched into one direction and hacking-free
mode while it is switched into the opposite direction. This method is simple, cost-
effective, and hack-proof. End users can use this method to have hacking-free
working environment without risking the firewall.
14.3 Related Prior Arts
Besides complicated networking settings and firewall software, a simple
hardware device was proposed by Fonseca (2003) by using a simple push/pull level
of a switch box to connect or disconnect the networking connection for the hacking
elimination. Fonseca called it as data line switch and filed for patent in the US on 24
July 2001. Later, Macuch (2005) designed the data line switch for the applications of
coaxial and DSL cables to control the computer connection to the Internet. Macuch
filed for a design patent in the USA on 17 November 2003. Of course, there is yet
another current practice by some end users to plug and unplug the networking cable.
242
Nevertheless, this method suffers from the hook damage of RJ45/RJ11 and
inconvenience access of networking port.
Here, a proposed component with similar function to data line switch is also
applicable to modular jack (aka modular connector) (Wikipedia Contributors, 2007m)
like RJ45 and RJ11. Modular connector was firstly invented by Hardesty (1975),
who filed it for patent in the US on 6 July 1973. RJ stands for registered jack
(Wikipedia Contributors, 20071). RJ45 and RJ11 are used as Ethernet jack and
telephone jack, respectively. Our new component is innovated from the dual in-line
package (DIL/DIP) switch by adding a collective actuator, which can be slide, rocker,
or piano-type (aka side/level).
The miniature DIP switch was found in the US patent database to be firstly
invented by Lockard (1977, 1979)， who filed for patent in the US on 25 March 1975
for the first time. Since then, there are various innovations on the DIP switch.
Hoffman (1982) had improved the manufacturing of DIP switch. Liataud and
Maloney (1983) had reduced the size, decreased the cost, and increased the reliability
of DIP switch. Brown (1983) had created the piano-type DIP switch. In the late
decade, Lin (1999) and Tai (2001) from Taiwan, R.O.C., had concomitantly
decreased the size, improved the manufacturing process, and increased the reliability
of DIP switch.
Normal slide switch wipes in parallel with the pin pairs. The slide actuator of
our proposed component wipes transversely to the pin pairs. The first slide switch
that can be found in the US patent database was invented by Bailey (1969). Even
though the wiping directions of normal slide switch and our switch are different by
90° degrees, their function is the same, i.e. to connect and disconnect the poles,
except the 10- way secure DIP switch oppositely switches two groups of poles.
14.4 Proposed Secure DIP Switch
For conventional n-way nPST (n Poles Single Throw) DIP switch, all the n
poles are independently switched on or off in parallel with the pin pairs 101 and 102.
A simple structural diagram of a 10-way DIP switch is shown in Figure 14.1. Here, a
modified DIP switch called secure DIP switch is innovatively proposed, where all
243
the individual switches of the DIP switch are joined and controlled simultaneously
by a transverse slider acting as an actuator in Figure 14.2. Alternative actuators are
raised/recessed slide, raised/recessed rocker, and piano-type (aka side/level). When a
USB connection is considered, an 8-way secure DIP switch for Ethernet cable will
become 10/ 12- way, or an extra 2/4- way secure DIP switch. The slider 103 can be
wiped transversely to the pin pairs 104 and 105 to either switch on the networking
connection and off the connection of the second storage device, or oppositely. This
means secure DIP switch is 10/ 12- way nPDT (n Poles Double Throws).
101
> 102
Figure 14.1 Structural diagram of conventional 10-way DIP switch
104
y io5
Figure 14.2 Structural diagram of proposed 10/12-way secure DIP switch
244
There are two groups of poles in opposite connections: 8-way RJ45/RJ1 1
networking connection and 2/4-way USB connection. 10Mbps and 100Mbps
Ethernet over twisted pair can use 4-way connection, but 1 Gbps/ 1 000Mbps Ethernet
must use 8-way connection. For USB connection, it can be 4 ways or 2 ways by
saving the power and ground cables. It is then integrated with two RJ45 sockets and
two USB sockets to form a simple and cost-effective innovation (Lee, 2008a, 2008b,
2008c).
If Category 5/5e cable (Wikipedia Contributors, 2007p) defined in
ANSI/TIA/EIA-568-A and TIA/EIA-568-B (Wikipedia Contributors, 2007k),
respectively, is used, the RJ45 socket will be backwards compatible with RJ11 for
two running pairs and one running pair, respectively.
14.5 Method and Device to Secure Hacking-Free Data Storage
Insofar as the secure DIP switch is specifically designed for a method and
device to secure a hacking-free data storage. An 8/10- way secure DIP switch is
integrated with two modular connector RJ45 sockets to connect or disconnect the
networking connection, and two optional USB sockets to oppositely disconnect or
connect the second storage device on a PCB (Printed Circuit Board). The integration
without USB sockets functioning as an RJ switch can be implemented as a wall plate
for new installation or as an external interconnection box for old design and
inconvenient switch access.
For the computer of the end user, a second storage device is needed. This can
be either an internal or external hard disk drive (HDD). It can also be a USB flash
drive. For external HDD and USB flash drive, they are hot-swappable when USB
port is used. For internal HDD of the type of SATA (Serial Advanced Technology
Attachment), a switch is needed to control the data connection. This switch called
HDD switch can be an 8-way secure DIP switch installed at the back panel of
computer with old design or at the front panel of computer with new design.
Similarly, the connection of external HDD and USB flash drive via USB port can
adopt a 2/4-way switch. This can get rid of the plug-and-play which can cause
reliability problem after frequent plugging and unplugging.
245
For a real implementation, an RJ switch is constructed as an interconnection
box from an 8-way secure DIP switch and two RJ45 sockets. The end user uses a
computer connected to an external HDD via USB port. The storage device can also
be a USB flash drive. There is an Ethernet cable connecting the computer and the
RJ45 socket of the interconnection box. Another Ethernet cable connects the second
RJ45 socket of the interconnection box and the networking wall plate. Clearly, these
can be easily understood by any normal end user. The 8-way switch can also be
made 10/12-way if the optional USB connection is added. Then, there are two
operating modes as in Table 14.1.
Table 14.1 Operating modes of method and device to secure hack-proof data storage
Operating Mode
Networking Connection
Second Storage Device
Hacking-free
Disconnected
Connected
Network access
Connected
Disconnected
For secure hacking-free operating mode, the actuator is switched to
disconnect the networking connection and then connect the second storage device.
The end user can create, open, modify, and store one's private and confidential data
in the second storage device. When network access is needed, the second storage
device is disconnected and then the network is connected. The end user can now surf
the Internet and one's data in the second storage device is safe from hacking via the
malicious computer network. Once the demand for network access has finished, the
end user can switch back to the hacking-free operating mode to manipulate the
private and confidential data.
14.6 Costs and Reliability
The current cost of a DIL switch in Malaysia ranges from MYR$3.88 to
MYR$46.77 depending on the contact ratings of voltage and current as well as the
operating life (Farnell, 2007). The FOREX (Foreign Exchange) of USD$1.00 was
about MYR$3.50 in September 2007 and October 2008. Mass purchase over 500
pieces can reduce the unit price of DIL switch to MYR$2.56. Subsequently, it can be
246
claimed that the added manufacturing cost is low and yet the added value of hacking-
free data storage is high.
The voltage and current of secure DIP switch will depend on the power over
the Ethernet cable (i.e. PoE (Power over Ethernet), PoL (Power over LAN) or Inline
Power) (Wikipedia Contributors, 2007o)， phone cable and USB connection.
Supplying power over Ethernet is strongly recommended to follow the IEEE
Standard. Clause 33 of "IEEE 802.3-2005 - Section Two" (LAN/MAN Standards
Committee, 2005) provides 48 volts DC over two of the four available pairs on a Cat.
3 / Cat. 5 cable with a maximum current of 400 mA for a maximum load power of
15.4 Watts (Wikipedia Contributors, 2007o).
For the Ethernet cable over LAN in Malaysia, it is normally Cat. 5 T568B
(Wikipedia Contributors, 2007k, 2007p). Contact rating of phone cable for network
usage is below the contact rating of Ethernet cable. If USB power cables travel
through the DIP switch, then it is 4 ways and the contact rating is 5.25 V DC and 500
mA. Otherwise, it needs 2 ways and the contact rating of USB data cables is below
2.8 V and 20 mA for high speed USB 2.0.
The reliability (aka operating life or service life) of DIP switch ranges from
1,000 to 35,000 operations. The death of DIP switch depends on the change of
contact resistance and the mechanical wear out of the actuator. It is expected that the
improvements by Lin (1999) and Tai (2001) can further increased the operating life
of DIP switch in parallel with the reduction of manufacturing materials, weight, and
cost. It is a question on the balance of costs and reliability.
This innovation is expected to be broadly used in the office environment,
where there exists a lot of private and confidential data. If the hacking-free operating
mode and network access operating mode are activated once a day for five times per
week, then the DIP switch can last for 3.85 years for the DIP switch with operating
life of 1,000 operations. The contact ratings, operating life, and cost of DIP switch
are closely correlated. Survey and research are needed for optimum manufacturing
design and supply chain management.
Yet another potentially broad application for men with good social status and
women with good conditions, this hack-proof data storage is also critical to protect
247
their human interaction network, daily itinerary, future plans, and financial accounts
from being maliciously conspired by the cronies of organized crime by using the
hacked secrets, flash mob approach, and sound snatching.
14.7 Other Forms of Innovation
An innovation of the improved 8-way 8PST DIL switch as in Figures 14.1-
14.2 is to become a 10- way 8PST+2PST DIL switch with an actuator activating
8PST and 2PST in opposite direction, where 8PST controls the network connection
of RJ-XX and 2PST is extendable to other nPST to control the hot-swappable USB
or SATA data/power connection to create a hack-proof data storage as in Figure 14.3.
As in Figures 14.1-14.2, the 8-way 8PST DIL switch acting as RJ switch for
wired Ethernet network can be modified to become 4-way 4PST DIL switch acting
as hot-swappable USB switch to control the wireless network connection using the
wireless USB network adapter operating on the wireless communication protocols
like Bluetooth, Wi-Fi, 3G， WiMAX, etc.
630
r
610
620
610
J
620
Figure 14.3 Innovated 10- way 8PST+2PST DIL switch activated in opposite
direction
248
In Figure 14.3， the 10- way 8PST+2PST DIL switch with reverse activation
630 can be modified to have the first 8PST 610 acting as RJ switch or to become
4PST acting as a USB switch for wireless USB network adapter in similarity with
Figures 14.1-14.2， and the second 2PST 620 is extendable to other nPST for other
types of data connection, like SATA and USB, to a storage device like HDD and
USB flash drive.
The improved DIL switch so far can be other types of switch performing
these enhanced functions to create hacking-free data storage, where they can also
switch on and off a few little switches to control the data and power connections like
keylock switch, selector switch, pushbutton switch, rocker switch, rotary switch,
slide switch, toggle switch, etc., with and without a light indicator of network
connection.
There are also some originally novel prototypes for this innovated DIP switch
in the forms of layout-design of integrated circuit in Malaysia (Lee, 2005b, 2006c,
2007b, 2007c, 2007d， 2007e， 2008d， 2008e， 2008f， 2008g).
14.8 Conclusion
Unless there is an advanced hacker who can interpret the weak
electromagnetic radiation across the secure DIP switch, this proposed method and
device for securing a hacking-free data storage can be claimed to be fully resisting
the hacking attacks. It is a simple integration consisting of a secure DIP switch, two
RJ45 sockets, and two optional USB sockets. The proposed switch adds little
manufacturing costs but highly added values, which may be a 10- way switch for a
RJ45/RJ11 and a USB connection. This hacking-free method and device is simple,
cost-effective, and hack-proof.
249
CHAPTER 15 CONCLUSIONS
15.1 Concise Summary
In a nutshell, this doctoral research project has contributed a lot of originally
novel knowledge contribution in the forms of methods, systems, and devices in the
fields of information engineering, generally, and security engineering, particularly.
The ways to evaluate a researcher is firstly discussed followed by the
essential condition to qualify for a doctoral degree. Contribution impact by referring
to the applications of research results for public usages is highly recommended.
Then in the first part, five methods and systems to create big and yet
memorizable secrets are presented in details. These five methods are as follows:
(i)
Chinese-character-encoded passphrase
(ii)
Two-dimensional (2D) key
(iii)
Multilingual key
(iv)
Multi-tier geo-image key
(v)
Multi-factor multimedia key using software token
(vi)
Hybrid combinations of any of the abovementioned five methods
Afterwards, the multimedia noises (or errors), enhanced frequency analysis,
information rate, and unicity distance are studied to show how to increase the
randomness of password/key to get higher entropy and securer protection.
To cater for the demands of multiple unique secrets to support various offline
and online accounts, the multihash key using the hash iteration and hash truncation is
invented here. To create more slave keys from a master key, three methods are
proposed: Using a filename, using a random number, and using a two -tier structure.
Later, there are three variants of multihash key to generate more slave keys. In
addition, multihash key is shown on how to act as a further authentication factor, as
well as simple key escrow method and system.
250
In the second part, applications of secret(s) and MePKC (Memorizable
Public-Key Cryptography) for twelve novel methods and systems are presented.
These applications show the future great contribution impacts that this doctoral
research project can trigger and cultivate. These twelve components are as follows:
(i) Memorizable symmetric key to resist quantum computer attack
(ii) Memorizable public-key cryptography (MePKC)
(iii) Other cryptographic, information- hiding, and non-cryptographic
applications of secret beyond 128 bits
(iv) Identification hardening of embedded data in steganography
(v) Electronic fund transfer using MePKC
(vi) Electronic software licensing using MePKC
^vn; MePKC human-computer and human- human authentication schemes
^vin) MePKC digital certificate having more than one asymmetric key pair
(ix) Three-tier MePKC digital certificates for ladder authentication
(x) Archiving the voice/ video calls of wired/wireless phones
(xi) Multipartite electronic commerce transactions using MePKC
(xuj Trust boosting of MePKC digital certificate by using more than one
certification authority and/or introducer of Trust of Web
Later, the MePKC digital timestamping scheme is proposed to act as the
digital evidence of copyright authorship and/or ownership to replace the SD
(Statutory Declaration) in Malaysia. For generalization, this MePKC digital
timestamping scheme can be applied as other IP (Intellectual Property) evidences,
especially the patent. After having this scheme, as well as knowing the great loss due
to the hacking and economic espionage, an innovated patent filing system called
"limited first-to-invent patent filing system" is proposed. This filing system is the
same as the first-to-invent patent filing system expect that there is a limited window
period from the first date of invention claim to the first patent filing date. This period
251
shall be about one to two years awaiting further research, survey, and evaluation. The
MePKC digital timestamping scheme shall be used to act as the IP evidence to prove
the first date of invention claim of this limited first-to-invent patent filing system.
Lastly, to secure the plaintext and decrypted ciphertext in the networked
computer system, the MePKC by itself is not enough due to the threat of virtual
hacking over the malicious and insecure computer communications network. Here, a
hack-proof (or hacking-free) data storage using an innovated DIP (Dual In-Line
Package) switch together with a second data storage device is proposed to secure the
plaintext and decrypted ciphertext. The files in the second data storage are always
offline whenever the computer has an Internet connection. When there is a need to
use the files in the second data storage, then the DIP switch has to switch off the
Internet connection first before activating the second data storage.
Let's create and maintain a networked info-computer age for a more
paperless, petroleum-less, and environment-friendly human society by having safer
multipartite electronic computer communications as from the original and novel
knowledge contribution of this research project.
15.2 Other Supporting Reading Materials in This Research Project
15.2.1 NIST Publication
NIST (National Institute of Standards and Technology), USA, has a lot of
publications on information engineering, generally, and security engineering,
particularly. These publications are very useful for further reading to discover and
cultivate novel and innovative ideas in applying the research outputs of this thesis.
The list of these NIST publications include personal identity verification (PIV)
card (Branstad, Clay & Hash, 2005; NIST, 2005a, 2005b, 2006b; Dray, Giles, Kelley
& Chandramouli, 2006; McCallister & Ferraiolo, 2006; MacGregor, Schwarzhoff &
Mehta, 2007; Polk, Dodson & Burr, 2007; Bailey, Chandramouli, Ghadiali &
Branstad, 2008)， discrete logarithm cryptography (Barker, Johnson & Smid, 2007)，
information security handbook (Bowen, Hash & Wilson, 2006; NIST, 2006d， 2007a,
2008; Singhal, Winograd & Scarfone, 2007; Tracy, Jansen, Scarfone & Winograd,
252
2007)， security requirements (NIST, 2001， 2004， 2006a, 2007d; Dent & Mitchell,
2004; Campbell & Easter, 2007a, 2007b, 2007c, 2008)， wireless security
(Karygiannis & Owens, 2002; Frankel, Eydt, Owens & Scarfone, 2007; Scarfone &
Dicoi, 2007)， SSL (Frankel, Hoffman, Orebaugh & Park, 2007)， RFID (Radio
Frequency Identification) (Karygiannis, Eydt, Barber, Bunn & Phillips, 2007)，
glossary of key information security terms (Kissel, 2006)， public-key cryptography
(Nechvatal, 1991)， computer authentication (NIST, 1985a, 1994b, 1997)， key
management (NIST, 1992), password (NIST, 1993), escrowed encryption (NIST,
1994a), computer security (NIST, 1995b), HMAC (Keyed-Hash Message
Authentication Code) (NIST, 2002a, 2007c), electronic mail (Email) (Tracy, Jansen,
Scarfone & Butterfield, 2007)， biometrics (Wilson, Grother & Chandramouli, 2007)，
and application- specific key management guidance (Barker, Burr, Jones, Polk, Rose
& Smid, 2008).
15.2.2 Security and Privacy
Besides the security, the entity privacy of an individual, family, and
organization has to be considered as well. Debra S. Herrmann (2007) has written a
book entitled "Complete Guide to Security and Privacy Metrics" for anyone would
like to know more about security and privacy in applying and implementing the
research outputs from this thesis.
15.2.3 Other Resources
There are also other resources of reading materials for interested readers to
know more. These include etymology dictionary of Malay language (Chong, 1997)
to know the relationship among various languages, font formats (Wikipedia
Contributors, 2008aj， 2008ak) for various styles to check the possibility to enlarge
the key space of multilingual key, computer (Stallings, 2006a), computer
communications network security (Stallings, 2000， 2005, 2006b, 2007);
cryptography (Koblitz, 1994; Nichols, 1999)， practical crypto system implementation
(Ferguson & Schneier, 2003)， open source network security tools (Schiffman, 2003)，
as well as finger reading (Lee, Tang, Chen & Fang, 2002) to know the existence of
253
ESP (Extra-Sensory Perception) in particular and psychic abilities in general
(Parapsycho logical Association, No date; Mitchell, 1974; Ostrander & Schroeder,
1974; Liu, 2001; Editors of Time-Life Books, 2004a, 2004b; Wikipedia Contributors,
2008bk, 2008bl， 2008bm， 2008bn， 2008bo， 2008bp) for the leaking possibilities of
key the secret.
15.3 Suggestions for Future Research
While reading the recommended supporting reading materials for this
research project, readers may also consider developing any of the suggested research
topics as discussed in this Section 15.3.
15.3.1 Fixed-Width Font Supporting All the Unicode Graphic Symbols
For current font file format, it can only support up to 65々3o characters or
graphic symbols. This is insufficient for multilingual key if all the Unicode graphic
symbols have to be included into a single font rile. Moreover, the future enlarged
font file to support all the Unicode graphic symbols has to be fixed-width font. This
is a potential IP (Intellectual Property) of design patent (aka industrial design) and
copyright to be developed in the coming future by interested person(s).
15.3.2 512-Bit Multihash Key Needs Hash Function beyond 1024 Bits
So far the popular and security intensively tested hash function is SHA
(Secure Hash Algorithm) family. The longest message digest of this SHA is SHA-
512 of SHA- 2 with 512 bits. This has limited the application of multihash key to
256-bit security for symmetric key and 128-bit security with 30-year protection for
asymmetric private key. To achieve the higher security strength at 256 bits of
symmetric key strength for 512-bit asymmetric private key, multihash key needs to
use 1024-bit hash function to generate 512-bit final slave key.
For 1024-bit hash function, there exists a scalable polymorphic hash function
(Roellgen, No date) to achieve this kind of message digest. Nevertheless, its security
254
strength is not well tested by the peer researchers in information security. Therefore,
while NIST is in the process of opening an website to accept the recommendation of
SHA-3, even though its maximum hash value requirement is 512 bits, related
researchers have to prepare themselves to go for a longer message digest up to 1024
bits to realize the 256-bit to 512-bit MePKC (Memorizable Public-Key
Cryptography).
15.3.3 MePKC Extension to Other Non-Conventional Cryptographic Schemes
In this thesis, the MePKC has been applied for encryption, digital signature,
authentication, digital cheque (aka electronic cheque), software licensing, public-key
certificate of public-key infrastructure (PKI), BAP (Byzantine Agreement Protocol),
electronic commerce, multihash signature, and digital timestamping.
Besides these conventional cryptographic schemes, interested researchers
may apply MePKC for other non-conventional cryptographic schemes like key
exchange, blind signature, multisignature, group-oriented signature, undeniable
signature, threshold signature, rail- stop signature, group signature, proxy signature,
signcryption, forward- secure signature, designated- verifier signature, copy protection,
electronic cash, electronic voting, MAC (Message Authentication Code), key escrow,
online verification of credit card, etc.
The blind signature scheme includes its further applications for electronic
cash (aka e-cash, electronic money, e-money, electronic currency, e-currency, digital
cash, digital money, digital currency, or scrip), and electronic voting (aka e-voting,
electronic election, e-election, electronic poll, e-poll, digital voting, digital election,
or digital poll).
15.3.4 Big Secret(s) for Information-Hiding and Non-Cryptographic
Applications
In addition to the big secret(s) applications for cryptographic schemes,
Section 9.4 has listed other applications of big secret(s) including the information
hiding and non-cryptographic applications. The information-hiding applications
255
include steganography, symmetric watermarking, and asymmetric watermarking.
The non-cryptographic applications are to be the seeds of PRNG (Pseudo-Random
Number Generator) and CSPRNG (Cryptographically Secure PRNG).
Hence, there are lots of spacious rooms to evaluate the key sizes and
corresponding bits of strength of these other applications of big secret(s). It is highly
expected for the existence of some literatures about their practically secure key
lengths and protection periods like the cryptographic schemes ("Cryptographic Key
Length Recommendation," No date; E. Barker, W. Barker, Burr, Polk, & Smid,
2007a, 2007b; Gehrmann & Naslund, 2005, 2006， 2007).
15.3.5 Safety Box Using Computerized Lock
For safety box using computerized lock (Domenicone, 2000)， its key pad is
purely numeric and the display panel is single-line. The short-term memory limits of
digits have been studied by Miller (195bj to be an average of 7 items plus or minus 2
(7 土 2) (Jones, 2002; Doumont, 2002)， and further studies show that they depends on
languages (Jones, 2002) in general and phonological short-term memory of 2- second
period (Baddeley, Thomson & Buchanan, 1975) in particular. It is 9.9 digits in
Chinese language (Hoosain & Salili, 1988) and 5.8 digits in Welsh language (Ellis &
Hennelly, 1980).
In other words, for single-line numeric passcode of this type of safety box, a
user using English, Chinese, or Welsh language will have a passcode with average
entropy of 23.25, 32.89， or 19.27 bits. The strength of these key lengths is insecure
whenever a brute force attack can be launched towards the safety box.
Therefore, 2-dimensional (2D) key is highly appreciated to be applied into
the safety box using computerized lock. For the key pad, it can remain to be purely
numeric in decimal digits or enlarged to become in hexadecimal digits.
15.3.6 Provable Security Studies
The only researcher, who is Kok-Wah Lee @ Xpree Jinhua Li, contributing
to the originally novel knowledge in this thesis, is educated in electrical engineering
256
in general and computer communications in particular. Hence, a lot of the proofs of
the inventions and innovations here are based on building up engineering prototypes.
Consequently, researchers in provable security, who are also mathematicians, are
expected to analyze thoroughly the security strength and loopholes of the algorithms,
methods, systems, devices, and apparatuses in security engineering as proposed in
this thesis.
15.3.7 Statistical Surveys for Various Security Schemes
Besides the provable security research over the inventions and innovations
proposed here, researchers in statistics can also consider conducting surveys like
some surveys (Adams & Sasse, 1999; Schneier, 2006; Florencio & Herley, 2007) to
know about the minimum, mean, maximum, and median key lengths of those
methods and systems to create big and yet memorizable secret as proposed here.
Similar statistical surveys can also be carried out for multihash key to know the
statistical values of master keys and slave keys.
15.4 Conclusions
To emphasize for the thrice time on the imperative aim of this research
project, here is the last paragraph.
Let's create and maintain a networked info-computer age for a more
paperless, petroleum-less, and environment-rnendly human society by having safer
multipartite electronic computer communications as from the original and novel
knowledge contribution of this research project.
257
to
ᄂ n
oo
Figure A. 1 Writing systems of the world
Reference: WiKipedia Contributors. (2008， August 27). Writing system, [Online]. Wikipedia the Free Encyclopedia.
Available: http://en.wiKipedia.Org/wiki/Image:WritingSystemsoftheWorld4.png [2008， September 1].
4
KIipM/l/lllMeCKIlll aJ({)clD[lT
ᄂ" |U -
-T^i^ — ^Sf^ ひ!
ュ』 _：：^^&
ᅮ ト. c
ね
字
« aiwc Art
^두
nnwvwaC ン I 대
t cwy-
*
リ aw
I 쑈 k
-, 에1、
wis A WAITING ^^Hi^ or^ Hfflw WOIiLD
Legend of writing systems of the world today:
Latin (alphabetic)
Cyrillic (alphabetic)
國 Hangul (featural alphabetic)
I Other alphabets
Arabic (abjad)
國 Other abjads
國 Devanagari (abugida)
Other abugidas
I Syllabaries
Chinese characters (logographic)
Table A. 1 Functional classification of writing systems
Type
Symbol Representation
Example
Pictographic
Pictorgram or iconic picture
Hieroglyph, Cuneiform
Ideographic
Ideogram
Way-finding sign, mathematical notation
Logographic
Morpheme
Chinese character
Syllabic
Syllable
Japanese kana
Alphabetic
Phoneme (consonant or vowel)
Latin alphabet
Abugida
Phoneme (consonant + vowel)
Indian Devanagari
Abjad
Phoneme (consonant)
Arabic alphabet
Featural
Phonetic feature
Korean hangul
259
Table A. 2 List of languages by number of native speakers
Language
Family
Ethnologue (Y2005)
1. Mandarin
^ m o- Tibetan Ch in e se
873,000,000
2. Hindi + Urdu
Indo-European, Indo- Iranian, Indo- Aryan
366,000,000
3. Spanish
Indo-European, Italic, Romance
358,000,000
4. English
Indo-European, Germanic, West
341,000,000
5. Arabic
Afro- Asiatic, Semitic
206,000,000
6. Portuguese
Indo-European, Italic, Romance
177,500,000
7. Ben
Indo-European, Indo- Iranian, Indo- Aryan
171,000,000
8 ᅵ Russian
Indo-European, Slavic, East
170,000,000
9. Japanese
J ap an e se-Ryukyu an
122,000,000
10. Cierman
Indo-European, Germanic, West
100,000,000
11. Punjabi
Indo-nuropean, Indo- Iranian, Indo- Aryan
88,000,000
12. French
Indo-European, Italic, Romance
79,572,000
13. Wu
Sino-Tibetan, Chinese
77,200,000
14. Javanese
Austronesian, Malayo-Polynesian, S 皿 da-Sulawesi
75,500,000
15. Korean
Considered either language isolate or Altaic
74,000,000
16. Telu2u
Dr a vidian South Central
69,700,000
17. Marathi
Indo-European, Indo- Iranian, Indo- Aryan
68,000,000
18. Vietnamese
Austro- Asiatic, M on - Khmer , Vietic
67,400,000
19. Tamil
Dra vidian, Southern
66,000,000
20. Italian
Indo-European, Italic, Romance
61,500,000
21. Cantonese
^ m o- Tibetan Ch in e se
54,800,000
22. Sindhi
Indo-European, Indo- Iranian, Indo- Aryan
54,500,000
23. Turkish
Altaic, Turkic, Oghuz
50,625,000
24. Min
5 m o- Tibetan Ch in e se
46,200,000
25. (Jujarati
Indo-European, Indo- Iranian, Indo- Aryan
46,100,000
26. Maithili
Indo-nuropean, Indo- Iranian, Indo- Aryan
45,000,000
27. Polish
Indo-European, Slavic, West
42,700,000
28. Ukrainian
Indo-European, Slavic, East
39,400,000
29. Persian
Indo-European, Indo-Iranian, Iranian
39,400,000
30. Malayalam
Dravidian, Southern - India
35,800,000
31. Kannada
Dravidian, Southern
35,400,000
32. Tamazight
Afro-Asiatic, Berber, Northern
32,300,000
Ref.: Wikipedia Contributors. (2008ad, July 22). List of languages by number of native speakers,
[Online]. Wikipedia the Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Lis ᄂ of_languages_by_number_of_native_speakers&
oldid=227300820 [2008, July 23].
260
APPENDIX B CHILD-MADE 2D KEYS
Authored by Wei-Dong Chui (徐伟 栋)， Wei-Jian Chui (徐伟 坚)， and Kok-Wah Lee (李 国华)
in January 2009
In this part, it is shown that children are also capable to create simple 2D keys by using the
key styles of ASCII art to draw some Chinese characters. The authors of these child-made 2D keys in
January 2009 in this Appendix B are 13-year-old Wei-Jian Chui born in 1996 (Figure B.l) and 9-year-
old Wei-Dong Chui born in 2000 (Figure B.2). To get the key size of every 2D key, just multiply the
number of ASCII characters of a 2D key by the value of 6.57 bits, or to be more accurate log 2 95. A
note here: Kok-Wah LEE being the main author has integrated each four Chinese characters created
by them to form a meaningful Chinese phrase for easy remembrance.
AAAAAA
AAAAAA
AVVVVA
AAAAAA
WWW
AAAAAA
AAAAAA
AAVVAA
AAVVAA
AAVVAA
WWW
AAVVAA
AAVVAA
AAVVAA
AAAAAA
AAAAAA
AAAAAA
WWW
AAAAAA
AAAAAA
AAAAAA
WWW
VAAAAV
VAAAAV
WWW
VAAAAV
VAAAAV
WWW
Two [二] Ten [十] One [一] Day [日]
Figure B.l 2D keys using ASCII art and Chinese characters meaning "twenty one days" [二^ 日]
AVVVVA
AAAAAA
WWW
AAVAAA
AVAAVA
WWW
AAAAAV
Cloud [云]
AVVVVA
AAAVAA
AAAVAA
WWW
AAVVAA
AVAAVA
VAAAAV
Sky [天]
AAAAAA
AAAAAA
WWW
AAVVAA
AAVVAA
WWW
AAAAAA
Job [ェ]
AAAVAAA
AAAVAAA
VVVVVVV
AAVVVAA
AVAVAVA
VAAVAAV
AAAVAAA
Wood [木]
Figure B.2 2D keys using ASCII art and Chinese characters meaning "cloudy sky nurtures the woods"
[云 天工 木]
261
REFERENCES
[I] 1790 Analytics Patent Analysis Intellectual Property Evaluation, [Online]. (No date).
Available: http://www.1790analytics.com [2008, September 3],
[2] Abadi, M.， Bharat, K.， and Marais, J. (2000, October 31). System and method for generating
unique passwords. USPTO Issued Patent US6141760, Filing Date: 31 October 1997， Issue
Date: 31 October 2000.
[3] Abadi, M.， Lomas, T. M. A., and Needham, R. (1997， December 16). Strengthening
passwords (Tech. Rep. No. SRC- 1997-033). Palo Alto, CA， USA: Hewlett-Packard
Company, HP Labs, Systems Research Center (SRC).
[4] Abadi, M.， Needham, R. M.， and Lomas, T. M. A. (2000, June 20). Method and apparatus
for strengthening passwords for protection of computer systems. USPTO Issued Patent
US6079021, Filing Date: 2 June 1997， Issue Date: 20 June 2000.
[5] Abril, P. S.， and Plant, R. (2007, January). The patent holder's dilemma: buy, sell, or troll?
Communications of the ACM, 50(1), 37-44.
[6] Adams, A., and Sasse, M. A. (1999， December). Users are not the enemy. Communications
of the ACM, 42(12), 41-46.
[7] Adams, A.， Sasse, M. A.， and Lunt, P. (1997， August 12-15). Making Passwords Secure and
Usable. Proceedings of the HCI on People and Computers XII, Bristol, UK, 1-19.
[8] Aguilera, M. K.， and Toueg, S. (1999， August 27). A simple bivalency proof that t-resilient
consensus requires t+1 rounds. Information Processing Letters, 77(3-4)， 155-158.
[9] Aharonian, G. (No date). Patent/copyright infringement lawsuits/licensing awards, [Online].
Source Translation and Optimization (STO). Available:
http://www.bustpatents.com/awards.htm [2007, October 10].
[10] Ahuja, V. (2000, May/June). Building trust in electronic commerce. IT Professional, 2(3),
61-63.
[II] Allan, A. (2004, December 6). Passwords are near the breaking point (Tech. Rep. No.
Gartner G00124970). Stamford, CT， USA: Gartner, Inc.
262
[12] Anderson, M.， Jaffe, R， Hibbert, C， Virkki, J.， Kravitz, J.， Chang, S., and Palmer, E. (2000,
February 1). Method and system for processing electronic documents. USPTO Issued Patent
US6021202, Filing Date: 19 December 1997, Issue Date: 1 February 2000.
[13] Anderson, M.， Jaffe, F.， Hibbert, C， Virkki, J.， Kravitz, J.， Chang, S., and Palmer, E. (2001,
March 27). Method and system for processing electronic documents. USPTO Issued Patent
US6209095, Filing Date: 31 August 1999， Issue Date: 27 March 2001.
[14] Anderson, R. (2001). Security engineering: A guide to building dependable distributed
systems. New York, NY, USA: John Wiley & Sons, Inc.
[15] Anderson, R. J.， and Petitcolas, F. A. P. (1998， May). On the limits of steganography. IEEE
Journal on Selected Areas in Communications, 7(5(4), 474-481.
[16] Arnold, M.， Schmucker, M.， and Wolthusen, S. D. (2003). Techinques and applications
digital watermarking and content protection. Norwood, MA, USA: Artech House, Inc.
[17] Aronson, E.， and Gerard, E. (1966， March). Beyond Parkinson's law: The effect of excess
time on subsequent performance. Journal of Personality and Social Psychology, 3(3), jj6-
339.
[18] Aronson, E.， and Landy, D. (1967， July). Further steps beyond Parkinson's Law: A
replication and extension of the excess time effect. Journal of Experimental Social
Psychology, 3(3), 274-285.
[19] Atreya, M.， Hammond, B.， Paine, S.， Starrett, P., and Wu, S. (2002). Digital signatures.
Berkeley, CA， USA: McGraw-Hill/Osborne.
[20] Baddeley, A. D.， Thomson, N., and Buchanan, M. (1975, December). Word length and the
structure of short-term memory. Journal of Verbal Learning and Verbal Behavior, 74(6),
575-589.
[21] Bailey, D.， Chandramouli, R.， Ghadiali, N.， and Branstad, D. (2008, February). Guidelines
for the accreditation of personal identity verification (PIV) card issuers (PCFs) (、draft)
(NIST Special Publication 800-79-1 (Draft)). Gaithersburg, MD， USA: CSRC (Computer
Security Resource Center), NIST.
263
[22] Bailey, J. R. (1969， December 23). Slide switch. USPTO Issued Patent US3485966, Filing
Date: 2 October 1968， Issue Date: 2 October 1968.
[23] Baltzley, C. A. (2000, November 28). Public key crypto system with roaming user capability.
USPTO Issued Patent US6154543, Filing Date: 25 November 1998， Issue Date: 28
November 2000.
[24] Baltzley, C. A. (2001a, August 16). Public key crypto system with roaming user capability.
USPTO Published Patent Application US2001/0014158, Filing Date: 28 March 2001.
[25] Baltzley, C. A. (2001b, September 18). Public key crypto system with roaming user capability.
USPTO Issued Patent US6292895, Filing Date: 19 June 2000, Issue Date: 18 September
2001.
[26] Barham, B.， and Foltz, J. (No date). University patenting and scientific advancement,
[Online]. Available: http://www.lafollette.wisc.edu/publicservice/stembarhamfoltz.ppt [2008,
September 3].
[27] Barker, E.， Barker, W.， Burr, W.， Polk, W.， and Smid, M. (2007a, March). Recommendation
for key management - Part 1: General (revised) (NIST Special Publication 800-57).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST, 61-71.
[28] Barker, E.， Barker, W.， Burr, W.， Polk, W.， and Smid, M. (2007b, March). Recommendation
for key management - Part 2: Best practices for key management organization (NIST
Special Publication 800-57). Gaithersburg, MD, USA: CSRC (Computer Security Resource
Center), NIST.
[29] Barker, E.， Burr, W., Jones, A.， Polk, T.， Rose, S., and Smid, M. (2008, August).
Recommendation for key management - Part 3: Application-specific key management
guidance (NIST Special Publication 800-57). Gaithersburg, MD, USA: CSRC (Computer
Security Resource Center), NIST.
[30] Barker, E.， Johnson, D.， and Smid, M. (2007, March). Recommendation for pair-wise key
establishment schemes using discrete logarithm cryptography (revised) (NIST Special
Publication 800-56A). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center),
NIST.
264
[31] Barker, E.， and Kelsey, J. (2007, March). Recommendation for random number generation
using deterministic random bit generators (revised) (NIST Special Publication 800-90).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[32] Barney, J. A., and Barney, J. R. (2003, April 29). Method and system for rating patents and
other intangible assets. USPTO Issued Patent US6556992, Filing Date: 14 September 2000,
Issue Date: 29 April 2003.
[33] Barni, M.， and Bartolini, F. (2004). Watermarking systems engineering: Enabling digital
assets security and other applications. New York, NY, USA: Marcel Dekker, Inc.
[34] Batista, P. D.， Campiteli, M. G.， and Kinouchi, O. (2006, July). Is it possible to compare
researchers with different scientific interests?. Sciento metrics, 68(1), 179-189.
[35] Bauer, F. L. (2002). Decrypted secrets: Methods and maxims of cryptology (3rd ed.). Munich,
Bavaria, Germany: Springer, 271-300.
[36] Bayh, B. (2006, December). Bayh-Dole: Don't turn back the clock. Les Nouvelles, [Online]
20 り <5(December)， 215-218. Licensing Executives Society International (LES I). Available:
http://www.lesi.org/BirchBayh/Bayh.pdf [2008, July 17].
[37] Beach, U. [2001, April 15]. How do we manage our expanding collections of passwords and
PINs?, [Online]. CIO.com. Available: http://www.cio.com/article/print/30161 [2008,
September 10].
[38] Beaver, K. (2004). Hacking for dummies. Indianapolis, Indiana, USA: Wiley Publishing, Inc.
[39] Behr, F. Jr., Fossum, V.， Mitzenmacher, M.， and Xiao, D. (2002). Estimating and comparing
entropies across written natural languages using PPM compression (Tech. Rep. No. TR-12-
02), [Online]. Cambridge, MA, USA: Harvard University, Harvard School of Engineering
and Applied Sciences. Available: ftp://ftp.deas.harvard.edu/techreports/tr-2002.html;
ftp://ftp.deas.harvard.edu/techreports/tr-12-02.ps.gz [2008, July 23].
[40] Behr, F.， Fossum, V., Mitzenmacher, M.， and Xiao, D. (2003, March 25-27). Estimating and
comparing entropies across written natural languages using PPM compression. Proceedings
of the Data Compression Conference 2003 (DCC 2003), Snowbird, Utah, USA, 416.
265
[41] Bellare, M.， Miner, S. K. (1999， August 15-19). A forward-secure digital signature scheme.
Proceedings of the 19th Annual International Cryptology Conference on Advances in
Cryptology 1999 (CRYPTO '99)， LNCS 1666， Santa Barbara, CA， USA, 431-448.
[42] Benaloh, J., and de Mare, M. (1991). Efficient broadcast time-stamping ( extended abstract),
[Online]. Technical Report 1， Department of Mathematics and Computer Science, Clarkson
University. CiteSeer. Available: http://citeseer.ist.psu.edu/benaloh91efficient.html [2007,
January 10].
[43] Berinato, S. (2007a, September 17). Hacker economics 1: Malware as a service - Who's
stealing your passwords? Global hackers create a new online crime economy, [Online].
CIO.com. Available: http://www.cio.com/article/print/135500 [2008, September 10].
[44] Berinato, S. (2007b, October 8). Hacker economics 2: The conspiracy of apathy, [Online].
CIO.com. Available: http://www.cio.com/article/print/135550 [2008, September 10].
[45] Berinato, S. (2007c, October 8). Hacker economics 3: MPACK and the next wave of malware.
[Online]. CIO.com. Available: http://www.cio.com/article/print/135551 [2008, September
10].
[46] Berinato, S. (2007d, October 8). Key malware terms - A layman 's glossary of malware terms,
[Online]. CIO.com. Available: http://www.cio.com/article/print/135453 [2008, September
10].
[47] Berinato, S. (2007e, October 8). How Gozi's first second unfolds, [Online]. CIO.com.
Available: http://www.cio.com/article/print/135451 [2008, September 10].
[48] Berinato, S. (2007f, October 8). Death by [Frame, [Online]. CIO.com. Available:
http://www.cio.com/article/print/135452 [2008, September 10].
[49] Bessen, J. (2004). Where have all the great inventors gonel, [Online]. Research on
Innovation. Available: http://www.researchoninnovation.org/GreatInventors.pdf [2008,
August 20].
[50] Beutelspacher, A. (1994). Cryptology (J. C. Fisher, Trans.). Washington, D.C., USA: The
Mathematical Association of America. (Original work published 1991 in German language).
266
[51] Bierly, P. E. III， Kolodinsky, R. W., and Charette, B. J. (No date). Understanding the
complex relationship between creativity and ethical ideologies. Journal of Business Ethics
[Online]. Available: http://dx.doi.org/10.1007/sl0551-008-9837-6 [2008, July 18].
[52] Bishop, M. (2003). Computer security: Art and science. Boston, MA, USA: Addison-Wesley.
[53] Blake, L， Seroussi, G.， and Smart, N. (1999， July). Elliptic curves in cryptography. In
London Mathematical Society Lecture Note Series (No. 265). Cambridge, Cambridgeshire,
UK: Cambridge University Press.
[54] Blake, I. R， Seroussi, G.， and Smart, N. P. (Eds.). (2005, April). Advances in elliptic curve
cryptography. In London Mathematical Society Lecture Note Series (No. 317). Cambridge,
UK: Cambridge University Press.
[55] Blaze, M.， Diffie, W., Rivest, R. ᄂ， Schneier, B.， Shimomura, T.， Thompson, E.， and Wiener,
M. (1996). Minimal key lengths for symmetric ciphers to provide adequate commercial
security. Chicago, Illinois, USA: A Report by an Ad Hoc Group of Cryptographers and
Computer Scientists.
[56] Blonder, G. (1996， September 24). Graphical password. USPTO Issued Patent US5559961,
Filing Date: 30 August 1995, Issue Date: 24 September 1996.
[57] Boatwright, M.， and Luo, X. (2007, September 28-29). What do we know about biometrics
authentication? ACM Proceedings of the 4th Annual Conference on Information Security
Curriculum Development 2007, Kennesaw, Georgia, USA, 205-209.
[58] Boneh, D.， and Franklin, M. (2006, September 26). Systems and methods for identity-based
encryption and related cryptographic techniques. USPTO Issued Patent US7 113594, Filing
Date: 13 August 2002, Issue Date: 26 September 2006.
[59] Bonnet, M.， Baroniunas, W.， and Webbink, M. (2008, August 28). Methods and systems for
tracking and auditing intellectual property in packages of open source software. USPTO
Published Patent Application US2008/0209399, Filing Date: 27 February 2007.
[60] Boren stein, N., and Freed, N. (1992, June). Base64 content-transfer-encoding. In Request for
comments ( 1341 ): MIME (Multipurpose Internet Mail Extensions ): Mechanisms for
specifying and describing the format of Internet message bodies (RFC 1341). Sterling,
Virginia, USA: Network Working Group, The Internet Engineering Task Force (IETF).
267
[61] Bose, B. K. (2006). Tips for getting a transactions article published, [Online]. IEEE Power
Electronics Society Newsletter, 2 りり (5(Second Quarter). Available: http://vlsi-
india.org/vsi/download/publications/archive/ieee-trans-tips.pdf [2008, August 2].
[62] Bo wen, P., Hash, J., and Wilson, M. (2006, October). Information security handbook: A
guide for managers (NIST Special Publication 800-100). Gaithersburg, MD， USA: CSRC
(Computer Security Resource Center), NIST.
[63] Boyd, C. (1989). Digital multisignatures, [Online]. In H. J. Beker, and F. C. Piper (Eds.),
Cryptography and Coding (pp. 241-246). Oxford, UK: Oxford University Press. Available:
http://sky.fit.qut.edu.au/~boydc/papers/ima89.pdf [2008, July 15].
[64] Brainard, J.， Juels, A.， Rivest, R. L.， Szydlo, M.， and Yung, M. (2006, October 30 -
November 3). Fourth-factor authentication: Somebody you know. Proceedings of the 13th
ACM Conference on Computer and Communications Security (CCS 2006), Alexandria,
Virginia, USA, 168-178.
[65] Branstad, D.， Clay, A., and Hash, J. (2005, July). Guidelines for the certification and
accreditation of PIV card issuing organizations (NIST Special Publication 800-79).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[66] Branstetter, L.， and Ogura, Y. (2005, August). Is academic science driving a surge in
industrial innovation? evidence from patent citations, [Online]. NBER Working Paper No.
11561. Available: http://www.nber.org/papers/wll561 [2008, September 16].
[67] Bray, M. J.， and Lee, J. N. (2000, September-November). University revenues from
technology transfer Licensing fees vs. equity positions. Journal of Business Venturing, 15(5-
6)， 385-392.
[68] Breitzman, A. (2006, April 13). Analysis of European patent referencing to IEEE papers,
conferences, and standards, [Online]. IEEE. Available:
http://www.ieee.org/portaVinnovate/freecontent/contentitems/whitepaper2.html [2008,
September 3].
[69] Breitzman, A. (2007, March 1^;. Analysis of patent referencing to IEEE papers, conferences,
and standards 1997-2006, [Online]. IEEE. Available:
http://wwwieee.org/portaycms_docs_iportals/iportals/publications/patentcitation/IEEB_Rep
ort_3-15-2006.pdf.pdf [2008， September 3].
268
[70] Breitzman, A. (2008, May 21). Analysis of patent referencing to IEEE papers, conferences,
and standards 1997-2007, [Online]. IEEE. Available:
http://www.ieee.org/portal/cms_docs_innovate/imiovate/freecontent/pdfs/IEEE_and_Patents
_5-21-08.pdf [2008, September 3].
[71] Breitzman, A. F.， and Narin, F. (2001, January 16). Method and apparatus for choosing a
stock portfolio, based on patent indicators. USPTO Issued Patent US6175824, Filing Date:
14 July 1999， Issue Date: 16 January 2001.
[72] Brown, A. S., Bracken, E.， Zoccoli, S., and Douglas, K. (2004, June 15). Generating and
remembering passwords. Applied Cognitive Psychology, 18(6), 641-651.
[73] Brown, P. F.， Pietra, V. J. D.， Mercer, R. L.， Pietra, S. A. D.， and Lai, J. C. (1982, March).
An estimate of an upper bound for the entropy of English. Computational Linguistics, 18(1),
31-40.
[74] Brown, R. P. (1983, June 21). Side actuated miniature DIP switch. USPTO Issued Patent
US4389549, Filing Date: 23 November 1981， Issue Date: 21 June 1983.
[75] Buck. W. (2000a). Mahabharata. New Delhi, India: Motilal Banarsidass.
[76] Buck. W. (2000b). Ramayana. New Delhi, India: Motilal Banarsidass.
[77] Bugaj, S. V. (1996， June). Passwords for real humans. USENIX Login, 27(3), 41.
[78] Burgess, C.， and Power, R. (2008, February 8). Secrets stolen, fortunes lost: Preventing
intellectual property theft and economic espionage in the 21st century. Burlington, MA, USA:
Syngress Publishing (now is at Elsevier' s Science & Technology publishing) (URL:
http://www.syngress.com) (URL: http://www.elsevierdirect.com).
[79] Burr, W. E.， Dodson, D. F.， and Polk, W. T. (2004, June). Electronic authentication
guideline (NIST Special Publication 800-63 Version 1.0). Gaithersburg, MD， USA: CSRC
(Computer Security Resource Center), NIST.
[80] Burr, W. E.， Dodson, D. R， and Polk, W. T. (2006, April). Electronic authentication
guideline (draft) (NIST Special Publication 800-63 Version 1.0.2 (Draft)). Gaithersburg, MD，
USA: CSRC (Computer Security Resource Center), NIST.
269
[81] Burr, W. E.， Dodson, D. F.， Perlner, R. A.， Polk, W. T.， Gupta, S., and Nabbus, E. A. (2008，
February 20). Electronic authentication guideline (NIST Special Publication 800-63-1).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[82] Cachin, C. (1998， April 14-17). An information-theoretic model for steganography.
Proceedings of the 2nd International Workshop on Information Hiding (IH '98), LNCS 1525,
Portland, Oregon, USA, 306-318.
[83] Callahan, D. W. (2001, August/September). The Ph.D. process. IEEE Potentials, 20(3\ 6-10.
[84] Cameron, K. (2005, May). The laws of identity, [Online]. Microsoft Corporation. Available:
http://msdn.microsoft.com/en-us/lihrary/ms996456.aspx [2008, August 25].
[85] Cameron, K.， and Jones, M. B. (2006, January). Design rationale behind the identity
met asy stem architecture, [Online]. Microsoft Corporation. Available:
http://research.miCTOsoft.com/~mbj/papers/Identity_Metasystem_Design_Rationale.pdf
[2008， August 25].
[86] Campbell, J., and Easter, R. J. (2007a, June 14). Security requirements for cryptographic
modules (Annex B): Approved protection profiles for FIPS PUB 140-2 (draft) (NIST FIPS
Pub 140-2 Annex B (Draft)). Gaithersburg, MD， USA: CSRC (Computer Security Resource
Center), NIST.
[87] Campbell, J., and Easter, R. J. (2007b, October 18). Security requirements for cryptographic
modules (Annex C): Approved random number generators for FIPS PUB 140-2 (draft) (NIST
FIPS Pub 140-2 Annex C (Draft)). Gaithersburg, MD， USA: CSRC (Computer Security
Resource Center), NIST.
[88] Campbell, J., and Easter, R. J. (2007c, December 18). Security requirements for
cryptographic modules (Annex A): Approved security functions for FIPS PUB 140-2 (draft)
(NIST FIPS Pub 140-2 Annex A (Draft)). Gaithersburg, MD， USA: CSRC (Computer
Security Resource Center), NIST.
[89] Campbell, J., and Easter, R. J. (2008, January 16). Security requirements for cryptographic
modules (Annex D): Approved key establishment techniques for FIPS PUB 140-2 (draft)
(NIST FIPS Pub 140-2 Annex D (Draft)). Gaithersburg, MD， USA: CSRC (Computer
Security Resource Center), NIST.
270
[90] Canning, P. (2006, May 22-25). Monetizing usage of scholarly collections with patent
citation analysis. Proceedings of the 27th IATUL Conference on Embedding Libraries in
Learning and Research (2006 IATUL), Porto, Norte, Portugal.
[91] CastleCops, [Online]. (No date). Available: http://www.castlecops.com [2008, September
25].
[92] Cavoukian, A. (2006, October). 7 laws of identity: The case for privacy-embedded laws of
identity in the digital age, [Online]. IPC (Office of the Information and Privacy
Commissioner) in Ontario, Canada. Available: http://www.ipc.on.ca/images/Resources/up-
71aws_whitepaper.pdf [2008, August 25].
[93] Cayre, R， Fontaine, C.， and Furon, T. (2005a, January 17-20). Watermarking security part
one: Theory. Proceedings of the SPIE Conference on Security, Steganography, and
Watermarking of Multimedia Contents VII, Vol. 5681, San Jose, CA, USA, 746-757.
[94] Cayre, R， Fontaine, C， and Furon, T. (2005b, January 17-20). Watermarking security part
two: Practice. Proceedings of the SPIE Conference on Security, Steganography, and
Watermarking of Multimedia Contents VII, Vol. 5681, San Jose, CA, USA, 758-768.
[95] Cayre, R， Fontaine, C.， and Furon, T. (2005c, October). Watermarking security: Theory and
practice. IEEE Transcations on Signal Processing, 53(10), 3976-3987.
[96] Chan dr an, P. M. (2007, August 7). Fully loaded - Realizing true potential of the intangible,
[Online]. IPFrontline Magazine of Intellectual Property and Technology, PatentCafe.com.
Available: http://www.ipfrontline.com/depts/article.asp ?id=15652&deptid=3 [2007, August
16].
[97] Chatterjee, S. (No date). Does increased equity ownership lead to more strategically involved
boards?. Journal of Business Ethics [Online]. Available: http://dx.doi.org/10.1007/sl0551-
008-9797-x [2008, July 20].
[98] Chaum, D. (1982, August 23-z5). Blind signatures for untraceable payments. Proceedings of
the 2nd Annual International Crypto logy Conference on Advances in Cryptology 1982 (aka
Proceedings of CRYPTO '82 on Advances in Cryptology) (CRYPTO ( 82), Santa Barbara, CA,
USA, 199-203.
[99] Chaum, D. L. (1988， July 19). Blind signature systems. USPTO Issued Patent US4759063,
Filing Date: 22 August 1983， Issue Date: 19 July 1988.
271
[100] Chaum, D.， and van Antwerpen, H. (1989, August 20-24). Undeniable signatures.
Proceedings of the 9th Annual International Crypto logy Conference on Advances in
Cryptology 1989 (CRYPTO ( 89), LNCS 435， Santa Barbara, CA， USA, 212-216.
[101] Chaum, D.， and van Heyst, E. (1991， April 8-11). Group signatures. Proceedings of the
Workshop on the Theory and Application of Cryptographic Techniques 1991 (aka
Proceedings of the Advances in Cryptology - Eurocrypt '91), LNCS 547, Brighton, UK, 257-
265.
[102] Chen, B. X. (2008, October 8). Google's super satellite captures first image, [Online].
Wired.com. Available: http://blog.wired.com/wiredscience/2008/10/geoeye-l-super.html
[2008, October 10].
[103] Chen, M. H. [陈明 华], (2004, November). 有组织 犯罪 问题对 策 研究 [Strategic study of
organized crime problems]. Beijing [：! 匕 iP、J， China [中国]: Publishing House of China
University of Political Science and Law [中 国政 法 大学 出 te 社], (in Chinese language).
[104] Chong, M. W. (1997). Ramus etimoiogi bahasa Melavu [Etymology dicuonary of Malay
language]. Shah Alam, Selangor, Malaysia: Penerbit Fajar Bakti Sdn. Bhd. (in Malay
language).
[105] CLJ Legal Network Sdn. Bhd. (1998). Solid Gold Publishers Sdn. Bhd. v. Chan Wee Ho &
Ors [1998] 5 CLJ /^5. In Law Database of Malaysia, [Online]. Available:
http://www.cljlaw.com [2008, May 20].
[106] Cohen, H.， and Frey, G. (2006). Handbook of elliptic and hyperelliptic curve cryptography.
Boca Raton, FL, USA: Taylor & Francis Group, Chapman & Hall/CRC
[107] Collberg, C. S., and Thomborson, C. (2002, August). Watermarking, tamper-proofmg, and
obfuscation - Tools for software protection. IEEE Transactions on Software Engineering,
25(8), 735-746.
[108] Comer, D. (Ed.). (1993). How to write a PhD dissertation or bedtime reading for people who
do not have time to sleep, [Online]. Available:
http://homes.cerias.purdue.edu/~spaf/Archive/dec.html [2008, August 2].
272
[109] Computing reviews, [Online]. (No date). ACM (Association for Computing Machinery).
Available: http://www.reviews.com [2008, August 21].
[110] Cooke, M. J., and Lebby, G. L. (1998, March 8-10). An optimal design for multilayer
feedforward networks. Proceedings of the 30th Southeastern Symposium on System Theory
1998 (SSST 1998), Morgantown, WV， USA, 507-511.
[Ill] Corell, S. (2000, May 1). Ten risks of PKI: In favour of smart card-based PKL Network
Security, 2000(5), 12-14.
[112] Coster, R. (2002, April). From first-to-invent to first-to-file: The Canadian experience,
[Online]. American Intellectual Property Law Association, Arlington, VA， US. Available:
http://www.torys.com/publications/pdf/ARTech-19T.pdf [2007, October 19].
[113] Cover, T. M.， and King, R. C. (1978， July). A convergent gambling estimate of the entropy
of English. IEEE Transactions on Information Theory, 24(4), 413-421.
[1 14] Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental
storage capacity. Behavioral and Brain Sciences [Online], 24(1), 87-185. Available:
http://www.bbsonline.Org/documents/a/00/00/04/46/index.html [2008， July 21].
[115] Cox, I. J.， Doerr, Q, and Furon, T. (2006, November 8-10). Watermarking is not
cryptography. Proceedings of the 5th International Workshop on Digital Watermarking 2006
(IWDW 2006), LNCS 4283, Jeju Island (aka Jejudo), Jeju Province (aka Jeju-do), South
Korea, 1-15.
[116] Cox, L， Miller, M.， Bloom, J., and Fridrich, J. (2007, November 16). Digital watermarking
and steganography (2nd ed.). San Fransisco, CA， USA: Morgan Kaufmann Publishers, Inc.
[117] Cryer, P. (2000). The research student's guide to success (2nd ed.)- Buckingham,
Buckinghamshire, UK: Open University Press.
[118] Crystal, D. (1999). The Penguin dictionary of language (2nd ed.). Middlesex, England:
Penguin Books.
[119] Cryptographic Key Length Recommendation, [Online]. (No date). Available:
http://www.keylength.com [2008, October 23].
273
[120] Curzon, L. B. (1998). Dictionary of law (4th ed.). Kuala Lumpur, Malaysia: International
Law Book Services.
[121] Dang, Q. (2007a, July 18). Randomized hashing digital signatures (draft) (NIST Special
Publication 800-106 (Draft)). Gaithersburg, MD， USA: CSRC (Computer Security Resource
Center), NIST.
[122] Dang, Q. (2007b, July 18). Recommendation for using approved hash algorithms (draft)
(NIST Special Publication 800-107 (Draft)). Gaithersburg, MD， USA: CSRC (Computer
Security Resource Center), NIST.
[123] Davies, J. H. E. (1997， March 4). Personal identification devices and access control systems.
USPTO Issued Patent US5608387, Filing Date: 26 May 1994， Issue Date: 4 March 1997.
[124] de Angeli, A.， Coventry, L.， Johnson, G., and Renaud, K. (2005, July). Is a picture really
worth a thousand words? Exploring the feasibility of graphical authentication systems.
International. Journal of Human- Computer Studies, 63(1-2), 128-152.
[125] de Koning Gans, G.， Hoepman, J.-H., and Garcia F. D. (2008, September 8-11). A practical
attack on the MIFARE Classic. Proceedings of the 8th IFIP WG 8.8/11.2 International
Conference on Smart Card Research and Advanced Applications (CARD IS 2008), London,
England, UK, 267-282.
[126] de Winter, B. (2008, October 7). Researchers show how to crack popular smart cards.
InfoWorld, [Online]. Available:
http://www.infoworld.com/article/08/10/07/Researchers_show_how_to_crack_popular_smart
_cards_l.html [2008， October 16].
[127] Deitel, H. M.， Deitel, P. J.， and Nieto, T. R. (2000). Internet and world wide web: How to
program. Upper Saddle River, New Jersey, USA: Prentice Hall.
[128] Deitel, H. M.， Deitel, P. J.， and Nieto, T. R. (2001). e-Business & e-commerce: How to
program. Upper Saddle River, New Jersey, USA: Prentice Hall.
[129] Dent, A. W., and Mitchell, C. J. (2004, October 31). User's guide to cryptography and
standards. Norwood, MA, USA: Artech House, Inc.
274
[130] desJardins, M. (1994， December). How to succeed in graduate school: A guide for students
and advisors (Part I of II). ACM Crossroads [Online], 7(2). Available:
http://www.acm.org/crossroads/xrdsl-2/advicel.html [2008, July 23].
[131] desJardins, M. (1995, February). How to succeed in graduate school: A guide for students
and advisors (Part II of II). ACM Crossroads [Online], 7(3). Available:
http://www.acm.org/crossroads/xrdsl-3/advice2.html [2008, July 23].
[132] Desmedt, Y. (1987， August 16-20). Society and group oriented cryptography: A new concept.
Proceedings of the Conference on the Theory and Applications of Cryptographic Techniques
1987 (aka Proceedings of the Advances in Crypto logy - CRYPTO ( 87) (aka Proceedings of
the 7th Annual International Crypto logy Conference on Advances in Cryptology 1987),
LNCS 293， Santa Barbara, CA， USA, 120-127.
[133] Desmedt, Y., and Frankel, Y. (1989， August 20-24). Threshold cryptosy stems. Proceeding of
the 9th Annual International Cryptology Conference on Advances in Cryptology 1989
(CRYPTO '89), LNCS 435, 307-315.
[134] DeVaney, D. B.， Harper, D. J.， and Short, T. D. (1987， October 6). Mobile telephone
switching office. USPTO Issued Patent US4698839, Filing Date: 3 June 1986， Issue Date: 6
October 1987.
[135] Devenport, M. (2005, April). Inventorship vs authorship: Who has the "write " to patent?,
[Online]. The Johns Hopkins University School of Medicine, Licensing and Technology
Development. Available: http://www.hopkinsmedicine.org/webnotes/licensing/0504.cfm
[2008, July 18].
[136] Dharmarajan, M. R. (2005, June 16). Method and apparatus for password generation.
USPTO Published Patent Application US2005/0 132203, Filing Date: 12 December 2003.
[137] Diffie, W.， and Hellman, M. E. (1976， November). New directions in cryptography. IEEE
Transaction on Information Theory, IT-22(6), 644-654.
[138] Diffie, W.， and Woods, W. A. (2006, June 22). Method for generating mnemonic random
passcodes. USPTO Published Patent Application US2007/0300076, Filing Date: 22 June
2006.
275
[139] Doggett, J.， Jaffe, F. A., and Anderson, M. M. (1997， October 14). Electronic funds transfer
instruments. USPTO Issued Patent US5677955, Filing Date: 7 April 1995, Issue Date: 14
October 1997.
[140] Domenicone, R. (2000, May 3). Safety box assembly, EPO Published Patent Application
EP0703341, Filing Date: 25 September 1995.
[141] Donner, I. H. (1999， December 7). Intellectual property audit system. USPTO Issued Patent
US5999907, Filing Date: 6 December 1993， Issue Date: 7 December 1999.
[142] Donner, I. H. (2001, July 17). Method of performing intellectual property (IP) audit
optionally over network architecture. USPTO Issued Patent US 62633 14, Filing Date: 3
March 2000, Issue Date: 17 July 2001.
[143] Doumont, J.-L. (2002, June). Magical numbers: the seven-plus-or-minus-two myth. IEEE
Transactions on Professional Communication, 45(2), 123-127.
[144] Dray, J. R， Giles, A., Kelley, M.， and Chandramouli, R. (2006, September). PIV card to
reader interoperability guidelines (NIST Special Publication 800-96). Gaithersburg, MD，
USA: CSRC (Computer Security Resource Center), NIST.
[145] Duane, M. J. (2008, Spring). Lending a hand: The need for public participation in patent
examination and beyond. Chicago-Kent Journal of Intellectual Property, 7(2), 57-74.
[146] Dudas, J. W. (2007, August 21). Changes to practice for continued examination filings,
patent applications containing patentably indistinct claims, and examination of claims in
patent applications (final rule), [Online]. US Federal Register Part II (37 CFR Part 1)，
72(161), 46715-46843. Department of Commerce, US: Patent and Trademark Office.
Available: http://www.ipfrontline.com/depts/article.asp ?id=15810&deptid=8;
http://www.ipfrontline.com/downloads/ContinuationRulesFinal.pdf [2007, August 22].
[147] Duren, M.， and Hosmer, C. (2002, August 6-9). Can digital evidence endure the test of time?.
Proceedings of the 2002 Digital Forensic s Research Workshop (DFRWS 2002 j, [Online],
Syracuse, New York, USA. Available: http://www.dfrws.org/2002/program.shtml [2007,
May 20].
[148] Eastlake, D. 3rd, Crocker, S., and Schiller, J. (1994， December). Request for comments
(1750): Randomness recommendations for security (RFC 1750). Sterling, Virginia, USA:
Network Working Group, The Internet Engineering Task Force (IETF).
276
[149] Eastlake, D. 3rd, and Hansen, T. (2006, July). Request for comments (4634): US Secure Hash
Algorithms (SHA and HMAC-SHA) (RFC 4634). Sterling, Virginia, USA: Network Working
Group, The Internet Engineering Task Force (IETF).
[150] Eastlake, D. 3rd, and Jones, P. (2001, September). Request for comments (3174): US Secure
Hash Algorithm 1 (SHA1) (RFC 3174). Sterling, Virginia, USA: Network Working Group,
The Internet Engineering Task Force (IETF).
[151] Ebert, L. B. (2004, Fall). Patent grant rates at the United States Patent and Trademark Office.
Chicago-Kent Journal of Intellectual Property, 4(1), 108-116.
[152] Ebert, L. B. (2005, Spring). Comment on "Patent grant rates at the United States Patent and
Trademark Office". Chicago-Kent Journal of Intellectual Property, 4(2), 186-195.
[153] Ebert, L. B. (2007, August 8). Patent grant rate lower than many academics think, [Online].
IPFrontline Magazine of Intellectual Property and Technology, PatentCafe.com. Available:
http://www.ipfrontline.com/depts/article.asp ?id=15684&deptid=4 [2007, August 16].
[154] Editors of Time-Life Books (Eds.)- (2004a). Mind over matter. London, England, UK:
Caxton Publishing Group.
[155] Editors of Time-Life Books (Eds.)- (2004b). The psychics. London, England, UK: Caxton
Publishing Group.
[156] Eggers, J. J., Su， J. K., and Girod, B. (2000). Asymmetric watermarking schemes, [Online].
Available: http://citeseer.ist.psu.edu/eggersOOasymmetric.html [2008, July 17].
[157] Elliott, D. R. (2007, September 11). Method for obtaining and allocating investment income
based on the capitalization of intellectual property. USPTO Issued Patent US7269566, Filing
Date: 20 August 2002, Issue Date: 11 September 2007.
[158] Ellis, N. C.， and Hennelly, R. A. (1980). A bilingual word-length effect: Implications for
intelligence testing and the relative ease of mental calculation in Welsh and English. British
Journal of Psychology, 71, 43-51.
[159] Ellison, C， and Schneier, B. (2000). Ten risks of PKI: What you're not being told about
public key infrastructure. Computer Security Journal, 7(5(1), 1-8.
277
[160] Elphinstone, L.， and Schweitzer, R. (1998). How to get a research degree: A survival guide.
St. Leonards, New South Wales, Australia: Allen & Unwin.
[161] Farnell. (2007). The Farnell inOne catalogue 2007/2008, [Book, Online]. Available:
http://my.farnell.com [2008, January 24].
[162] Feistel, H. (1974a, March 19). Block cipher cryptographic system. USPTO Issued Patent
US3798359, Filing Date: 30 June 1971， Issue Date: 19 March 1974.
[163] Feistel, H. (1974b, March 19). Centralized verification system. USPTO Issued Patent
US3798360, Filing Date: 30 June 1971， Issue Date: 19 March 1974.
[164] Feistel, H. (1974c, March 19). Step code ciphering system. USPTO Issued Patent US3798605,
Filing Date: 30 June 1971， Issue Date: 19 March 1974.
[165] Ferguson, N.， Schneier, B. (2003). Practical cryptography. Indianapolis, Indiana, USA:
Wiley Publishing, Inc.
[166] Fialka, J. J. (1999， March 1). War by other means: Economic espionage in America. New
York, NY, USA: W. W. Norton & Company (URL: http://www.wwnorton.com).
[167] Finegan, E. (2004). Language: Its structure and use (4th ed.). Boston, MA: Thomson
Wardsworth.
[168] Fink, S. (2002, January 15). Sticky fingers: Managing the global risk of economic espionage
(hardcover). New York, NY, USA: Dearborn Trade Publishing (now called as Kaplan
Publishing) (URL: http://www.dearborntrade.com) (URL: http://www.kaplanpublishing.com).
[169] Fink, S. (2003, December 9). Sticky fingers: Managing the global risk of economic espionage
(paperback). Bloomington, IN, USA: iUniverse.com, a self-publishing company.
[170] Fischer, M.， and Lynch, N. (1982, June 13). A lower bound for the time to assure interactive
consistency. Information Processing Letters, 14(4), 183-186.
[171] Fitzi, M.， and Hirt, M. (2006, July 23-26). Optimally efficient multi-valued Byzantine
agreement. Proceedings of the 2^th Annual ACM Symposium on Principles of Distributed
Computing 2006 (PODC 2006), Denver, Colorado, USA, 163-168.
278
[172] Florencio, D.， and Herley, C. (2007, May 8-12). A large-scale study of web password habits.
Proceedings of the 16th ACM International Conference on World Wide Web, Banff, Alberta,
Canada, 657-666.
[173] Fonseca, D. E. (2003， December 9). Data line switch. USPTO Issued Patent US6660950,
Filing Date: 24 July 2001, Issue Date: 9 December 2003.
[174] Ford, W. (1994). Computer communications security: Principles, standard protocols and
techniques. Englewood Cliffs, NJ, USA: Prentice Hall.
[175] Ford, W.， and Baum, M. S. (2001). Secure electronic commerce: Building the infrastructure
for digital signatures and encryption (2nd ed.)- Upper Saddle River, New Jersey, USA:
Prentice Hall.
[176] Fossum, V. (No date). Entropy, compression, and information content, [Online]. Available:
http://www.eecs.umich.edu/~vfossum/pubs/entropy_explanation.pdf;
http://www.cs.iupui.edu/~xkzou/teaching/csci590/entropy_explanation.pdf [2008, July 23].
[177] Frankel, S.， Eydt, B.， Owens, L.， and Scarf one, K. (2007, February). Establishing wireless
robust security networks: A guide to IEEE 802.1 li (NIST Special Publication 800-97).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[178] Frankel, S.， Hoffman, P., Orebaugh, A.， and Park, R. (2007, August). Guide to SSL VPNs
(draft) (NIST Special Publication 800-113 (Draft)). Gaithersburg, MD， USA: CSRC
(Computer Security Resource Center), NIST.
[179] Franks, J., Hallam-Baker, P., Hostetler, J.， Lawrence, S.， Leach, P., Luotonen, A.， and
Stewart, L. (1999， June). Request for comments (2617): HTTP authentication: Basic and
digest access authentication (RFC 2617). Sterling, Virginia, USA: Network Working Group,
The Internet Engineering Task Force (IETF).
[180] Fridrich, J.， and Goljan, M. (2004, December 14). Reliable detection of LSB steganography
in color and grayscale images. USPTO Issued Patent US6831991, Filing Date: 22 June 2001,
Issued Date: 14 December 2004.
[181] Fridrich, J.， Goljan, M.， and Soukal, D. (2004, January 18-22). Searching for the stego-key.
Proceedings of the SPIE on Security, Steganography, and Watermarking of Multimedia
Contents VI, San Jose, California, USA, 70-82.
279
[182] Fried, L， and Evers, J. (2006, February 14). Gates: End to passwords in sight, [Online].
CNET Networks, Inc. Available: http://news.cnet.com/Gates-End-to-passwords-in-
sight/2100-7355_3-6039177.html?tag=nw. 1 1 [2008, August 23].
[183] Fumy, W., and Landrock, P. (1993， June). Principles of key management. IEEE Journal on
Selected Areas in Communications, 77(5), 785-793.
[184] Furht, B.， and Kirovski, D. (2006a, April 18). Multimedia watermarking techniques and
applications. Boca Raton, FL， USA: Taylor & Francis Group, Auerbach Publications.
[185] Furht, B.， and Kirovski, D. (2006b, May 3). Multimedia encryption and authentication
techniques and applications. Boca Raton, FL， US A: Taylor & Francis Group, Auerbach
Publications.
[186] Furnell, S. (2005, March). Authenticating ourselves: will we ever escape the password?
Network Security, 2005(3), 8-13.
[187] Furon, T.， and Duhamel, P. (2003, April). An asymmetric watermarking method. IEEE
Transactions on Signal Processing, 57(4), 981-995.
[188] Furon, T. (2005, September 15-17). A survey of watermarking security. Proceedings of the
4th International Workshop on Digital Watermarking 2005 (IWDW 2005), LNCS 3710,
Siena, Tuscany, Italy, 201-215.
[189] Gabber, E.， Gibbons, P. B.， Matias, Y.， and Mayer, A. (1997， February 24-28). How to make
personalized web browsing simple, secure, and anonymous. Proceedings of 1st International
Conference on Financial Cryptography 1997 (FC 1997), LNCS 1318, Anguilla, British West
Indies, 17-31.
[190] Ganesan, R. (1996a, July 9). Yaksha, an improved system and method for securing
communications using split private key asymmetric cryptography. USPTO Issued Patent
US5535276, Filing Date: 9 November 1994， Issue Date: 9 July 1996.
[191] Ganesan, R. (1996b, September 17). System and method for centralized session key
distribution, privacy enhanced messaging and information distribution using a split private
key public cryptosystem. USPTO Issued Patent US5557678, Filing Date: 18 July 1994， Issue
Date: 17 September 1996.
280
[192] Ganesan, R. (1998a, April /). Computer system for securing communications using split
private key asymmetric cryptography. USPTO Issued Patent US5737419, Filing Date: 7 June
1996， Issue Date: 7 April 1998).
[193] Ganesan, R. (1998b, May 5). Securing E-mail communications and encrypted file storage
using yaksha split private key asymmetric cryptography. USPTO Issued Patent US5748735,
Filing Date: 7 June 1996， Issue Date: 5 May 1998.
[194] Ganesan, R. (1998c, November 17). Computer system for centralized session key distribution,
privacy enhanced messaging and information distribution using a split private key public
cryptosystem. USPTO Issued Patent US5838792, Filing Date: 8 August 1996， Issue Date: 17
November 1998.
[195] Ganesan, R. (1999， May 18). Programmed computer for identity verification, forming joint
signatures and session key agreement in an RSA public cryptosystem. USPTO Issued Patent
US5905799, Filing Date: 15 October 1996， Issue Date: 18 May 1999.
[196] Ganesan, R.， Sandhu, R. S.， Cottrell, A. P., and Austin, K. (2006a, May 31). Augmented
single factor split key asymmetric cryptography-key generation and distributor. USPTO
Published Patent Application US2007/0033392, Filing Date: 31 May 2006.
[197] Ganesan, R.， Sandhu, R. S., Cottrell, A. P., and Austin, K. (2006b, May 31). Secure login
using augmented single factor split key asymmetric cryptography. USPTO Published Patent
Application US2007/0 186095, Filing Date: 31 May 2006.
[198] Ganesan, R.， Sandhu, R. S., Cottrell, A. P., and Austin, K. (2006c, May 31). Secure login
using single factor split key asymmetric cryptography and an augmenting factor. USPTO
Published Patent Application US2007/0033393, Filing Date: 31 May 2006.
[199] Ganesan, R.， Sandhu, R. S., Cottrell, A. P., Schoppert, B. J.， and Bellare, M. (2006, May 2).
Protecting one-time-passwords against man- in-the- middle attacks. USPTO Published Patent
Application, US2007/0033642, Filing Date: 2 May 2006.
[200] Ganesan, R.， and Yacobi, Y. (1996, December 24). System and method for identity
verification, forming joint signatures and session key agreement in an RSA public
cryptosystem. USPTO Issued Patent US5588061, Filing Date: 20 July 1994， Issue Date: 24
December 1996.
281
[201] Garcia, F. D.， de Koning Gans, G., Muijrers, R.， van Rossum, P., Verdult, R.， Schreur, R. W.,
and Jacobs, B. (2008, October 6-8). Dismantling MIFARE Classic. Proceedings of the 13th
European Symposium on Research in Computer Security (ESORICS 2008), Malaga,
Andalusia, Spain, 97-114.
[202] Gardiner, D. (2008). Phishing ativity tends rport Ql/2008, [Online]. APWG (Anti-Phishing
Working Group). Available: http://www.antiphishing.org/reports/apwg_report_Ql_2008.pdf
[2008, September 19].
[203] Gehringer, E. F. (2002, June 6-8). Choosing passwords: Security and human factors.
Proceedings of the IEEE International Symposium on Technology and Society (1ST AS 2002),
Raleigh, North Carolina, USA, 369-373.
[204] Gehrmann, C， and Naslund, M. (Eds.). (2005, March 1). ECRYPT Yearly report on
algorithms and keysizes (2004) (Report No. IST-2002-507932 D.SPA.10). Katholieke
Universiteit Leuven, Leuven-Heverlee, Belgium: European Network of Excellence in
Cryptology (ECRYPT).
[205] Gehrmann, C， and Naslund, M. (Eds.). (2006, January 29). ECRYPT Yearly report on
algorithms and keysizes (2005) (Report No. IST-2002-507932 D.SPA.21). Katholieke
Universiteit Leuven, Leuven-Heverlee, Belgium: European Network of Excellence in
Cryptology (ECRYPT).
[206] Gehrmann, C.， and Naslund, M. (Eds.). (2007, January 26). ECRYPT Yearly report on
algorithms and keysizes (2006) (Report No. IST-2002-507932 D.SPA.16). Katholieke
Universiteit Leuven, Leuven-Heverlee, Belgium: European Network of Excellence in
Cryptology (ECRYPT).
[207] Gennaro, R. (2005, April). An improved pseudo-random generator based on the discrete
logarithm problem. Journal of Cryptology, 18(2), 91-110.
[208] Glick, L. (1995). Criminology. Needham Heights, MA, USA: Allyn & Bacon.
[209] Gladden, C. A., and Parelman, M. H. (1979， May 1). Rapidly deploy able emergency
communication system. USPTO Issued Patent US4 152647, Filing Date: 23 February 1978,
Issue Date: 1 May 1979.
282
[210] Goal, P. M.， and Kriese, S. J. (2004, August 26). Method and system for automated password
generation. USPTO Published Patent Application US2004/0168068, Filing Date: 20
February 2003.
[211] Goldstein, H. (2006, November). Patent power. IEEE Spectrum, 43(11)， 44-47.
[212] Goldwasser, S. (1997， October 20-22). New directions in cryptography: twenty some years
later (or cryptograpy and complexity theory: a match made in heaven). Proceedings of the
38th Annual Symposium on Foundations of Computer Science (SFCS 1997), Miami Beach,
Fl， USA, 314-324.
[213] Good practices in the protection of witnesses in criminal proceedings involving organized
crime, [Online]. (2008). UNODC (United Nations Office on Drugs and Crime). Available:
http://www.unodc.org/documents/organized-crimeAVitness-protection-manual-Feb08.pdf
[2008, July 21].
[214] Gordon, R. G. Jr. (Ed.)- (2005, January 1). Ethnologue: Languages of the world (15th ed.)，
[Online]. Dallas, Texas, USA: SIL International. Available:
http://www.ethnologue.com/web.asp [2008, September 15].
[215] Gouda, M. G., Liu, A. X.， Leung, L. M.， and Alam, M. A. (2005， June 7-10). Single
password, multiple accounts. Proceedings of the 3rd International Conference on Applied
Cryptography and Network Security (ACNS 2005), LNCS 3531， New York City, NY, USA,
Industry I Short Paper Track, 1-12.
[216] Graefe, J. M.， Lashley, L.， Guimaraes, M. A. M.， Guodabia, E.， Gupta, A. K.， Henry, H.， and
Austin, R. (2007, September 28-29). Credit card transaction security. ACM Proceedings of
the 4th Annual Conference on Information Security Curriculum Development 2007,
Kennesaw, Georgia, USA, 118-123.
[217] Haber, S.， and Stornetta, W. S. (1990， August 11-15). How to time- stamp a digital document.
Proceedings of the 10th Annual International Cryptology Conference on Advances in
Cryptology 1990 (CRYPTO LNCS 537, Santa Barbara, CA， USA, 437-455.
[218] Haber, S.， and Stornetta, W. S. (1991， January). How to time- stamp a digital document.
Journal of Cryptology, 3(2), 99-111.
[219] Hachez, G.， and Quisquater, J.-J. (2002). Which directions for asymmetric watermarking?,
[Online]. Available: http://citeseer.ist.psu.edu/564734.html [2008, July 17].
283
[220] Halderman, J. A., Waters, B.， and Felten, E. W. (2005, May 10-14). A convenient method for
securely managing passwords. Proceedings of the 14th International Conference on World
Wide Web 2005, Chiba, Japan, 471-479.
[221] Hanker son, D.， Menezes, A., and Van stone, S. (2004). Guide to elliptic curve cryptography.
New York, NY, USA: Springer- Verlag New York, Inc.
[222] Hanker son, D.， Menezes, A.， and Van stone, S. (2005, August). Guide to elliptic curve
cryptography [摘 圆曲线 密码学 导论] (张焕 国， Trans, to Chinese language). Beijing [北京]，
China [中国]: Publishing House of Electronics Industry [电 子ェ业 出版社]. (Original
published in 2004 in English language).
[223] Haperen, P. V. (1997， May 16). Graphical password entry. UK Published Patent Application
GB23 13460, Filing Date: 16 May 1997.
[224] Hardesty, E. C. (1975, January 14). Electrical connecting devices for terminating cords and
methods of assembling the devices to cords. USPTO Issued Patent US 3 8603 16, Filing Date:
6 July 1973， Issue Date: 14 January 1975.
[225] Harn, L.， and Kresler, T. (1989, July 20). New scheme for digital multisignatures.
Electronics Letters, 25(15), 1002-1003.
[226] Hart, G. W. (1994， September). To decode short cryptograms. Communications of the ACM,
37(9), 102-108.
[227] Hartung, R， and Kutter, M. (1999， July). Multimedia watermarking techniques. Proceedings
of the IEEE, 87(1), 1079-1107.
[228] Haskin, J. (2008, January 11). Busting the 10 myths about data protection, [Online].
CIO.com. Available: http://www.cio.com/article/print/171551 [2008, September 10].
[229] Hayes, J. (2001). Fiber optics technician's manual (2nd ed.). Albany, NY, USA: Thomson
Learning, Inc., Delmar.
[230] Haykin, S. (1999). Neural networks: A comprehensive foundation (2nd ed.). Upper Saddle
River, NJ， USA: Prentice Hall.
284
[231] Haylock, S. E. (No date). Fingerprints, [Online]. Answers.com. Available:
http://www.answers.com/topic/fingerprints-5 [2008, September 1].
[232] He, J. H. [何 家弘] (Ed.). (2007， August). 新编 犯罪 侦查学 [New criminal investigation
study]. Beijing [北京]， China [中国]: Publishing House of China's Legal System [中国 法制
出版社], (in Chinese language).
[233] Herrmann, D. S. (2007, January 22). Complete guide to security and privacy metrics:
Measuring regulatory compliance, operational resilience, and ROI. Boca Raton, Florida,
USA: Auerbach Publications, Taylor & Francis Group.
[234] Hilley, S. (2006, October). Internet war: Picking on the finance sector - Survey More
vulnerabilities & phishing. Computer Fraud & Security, 2006(10), 2.
[235] Hirsch, J. E. (2005, November 15). An index to quantify an individual's scientific research
output. PNAS (Proceedings of the National Academy of Sciences), ᅵ /ひ 2(46)， 16569-165 /z.
[236] Hoffman, N. E. (1982, June 1). DIP swtich. USPTO Issued Patent US4332987, Filing Date:
15 December 1980， Issue Date: 1 June 1982.
[237] Hoosain, R.， and Salili, F. (1988). Language differences, working memory, and mathematical
ability. Practical aspects of memory: Current research and issues, 2, 512-517.
[238] Huang, T. D. (1985, February 19). Method for encoding Chinese characters. USPTO Issued
Patent US4500872, Filing Date: 18 March 1982, Issue Date: 19 February 1985.
[239] Huang, X. R. [黄秀 如]. (2002). 词典 的两个 世界 [A history of dictionaries]. Taipei,
Taiwan (ROC): Net and Books [网路 与书] . (in Chinese language).
[240] Hunter, P. (2006, May). Microsoft declares war on phishers. Computer Fraud & Security,
2006(5)， 15-16.
[241] Identity Theft Resource Center (ITRC), [Online]. (No date). Available:
http://www.idtheftcenter.org [2008, October 23].
[242] IEEE Communications Society. (2002). A brief history of communications. Piscataway, NJ:
IEEE Communications Society.
285
[243] IEEE Publications. (2002, February 15 - 2008, June 20). IEEE Publication Services and
Products Board (PSPB) Operations Manual, [Online]. IEEE Publications, Piscataway, NJ,
USA. Available:
http://wwwieee.org/portaycms_docs_iportals/iportals/publications/PSPB/opsmanual.pdf
[2008, July 18].
[244] Identity-related crime. (2007, November 29-30). In UNODC and organized crime, [Online].
UNODC (United Nations Office on Drugs and Crime). Available:
http://www.unodc.org/unodc/en/organized-crime/index.html [2008, July 21].
[245] IIPA. (2001). Rock Records (M) Sdn. Bhd. v. Audio One Entertainment Sdn. Bhd. [2005] 1
CLJ 200. In Special report 301 Malaysia, [Online]. International Intellectual Property
Alliance (IIPA). Available: http://www.iipa.com/rbc/2001/2001SPEC301MALAYSIA.pdf
[2007， May 20].
[246] ILBS. (2003, March 20). Laws of Malaysia: Evidence Act 1950 (Act 56). Petaling Jaya,
Selangor, Malaysia: International Law Book Services (ILBS).
[247] International Dunhuang Project (IDP) [國際 敦煌 項目]. (No date). The International
Dunhuang Project: The Silk Road online, [Online]. Available: http://idp.bl.uk (British
Library); http://idp.nlc.gov.cn (National Library of China) [2008, July 16].
[248] Inventors Assistance League. (No date). First to invent vs. first to file, [Online]. Available:
http://www.inventions.org/resources/advisory/first.html [2007, October 22].
[249] Inventorship vs. authorship, [Online]. (2008). Albert Einstein College of Medicine of
Yeshiva University, Offices of Biotechnology and Business Development. Available:
http://www.aecom.yu.edu/biotechnology/page.aspx ?id=3316 [2008, July 18].
[250] Ismail, S. (2001). Time -stamping for Malaysia, [Online]. CiteSeer. Available:
http://citeseer.ist.psu.edu/ismail01timestamping.html [2007, January 8].
[251] Itakura, K.， and Nakamura, K. (1983， October). A public key cryptosystem suitable for
digital multisignatures. NEC Journal on Research & Development, 77， 1-8.
[252] Ives, B.， Walsh, K. R.， Schneider, H. (2004, April). The domino effect of password reuse.
Communications of the ACM ， 47(4), 73-/8.
286
[253] Jablon, D. P. (2006, March 7). Cryptographic methods for remote authentication. USPTO
Issued Patent US7010692, Filing Date: 9 June 2004, Issue Date: 7 March 2006.
[254] Jacobs, A. (1998). Dictionary of music (6th ed.). England, United Kingdom: Penguin Books.
[255] Jacobs, R. A. (1988). Increased rates of convergence through learning rate adaption. Neural
Networks, 7(4), 295-307.
[256] Jan sen, W.， Gavrila, S., Korolev, V.， Ayers, R.， and Swanstrom, R. (2003, July). Picture
password: A visual login technique for mobile devices (NIST IR 7030). Gaithersburg, MD，
USA: CSRC (Computer Security Resource Center), NIST.
[257] Jermyn, I.， Mayer, A., Monrose, F.， Reiter, M.， and Rubin, A. (1999， August 23-26). The
design and analysis of graphical passwords. Proceedings of 8th USENIX Security Symposium.
Washington D.C., USA, 1-14.
[258] Jones, D. M. (2002, October). The 7±2 urban legend, [Online]. MISRA C Conference.
Available: http://citeseer.ist.psu.edu/jones02urban.html;
http://www.knosof.co.uk/cbook/misart.pdf [2008, July 20].
[259] Jones, M. B. (2005, May). Microsoft's vision for an identity metasystem, [Online]. Microsoft
Corporation. Available: http://msdn.microsoft.com/en-us/lihrary/ms996422.aspx [2008,
August 25].
[260] Kahn, D. (1996a, May 30 - June 1). The history of steganography. Proceedings of the 1st
International Workshop on Information Hiding (IH '96)， LNCS 1174, Cambridge,
Cambridgeshire, UK, 1-5.
[261] Kahn, D. (1996b, December 5). The code -breakers: The comprehensive history of secret
communication from ancient times to the Internet (Revised and updated ed.)- New York City,
NY, USA: Scribner, 93-105, 125.
[262] Kanaley, R. (2001, February 4). Login error trouble keeping track of all your sign-ons?
Here へ a place to keep your electronic keys, but you'd better remember the password. San
Jose Mercury News.
[263] Kang, M. M.， Park, W. W., and Koo, J. R. (2003， June 28 - July 6). Agent for electronic
commerce on the semantic web. Proceedings of the 7th Korea-Russia International
287
Symposium on Science and Technology 2003 (KORUS 2003 J, Ulsan, Republic of Korea,
2003， Vol. 2， 360-363.
[264] Karp, A. H. (2003， May 6). Site-specific passwords (Tech. Rep. No. HPL-2002-39R1). Palo
Alto, CA， USA: Hewlett-Packard Company, HP Laboratories Palo Alto, Intelligent
Enterprise Technologies Laboratory.
[265] Karp, A. H.， and Poe, D. T. (2002, August 2). System-specific passwords. USPTO Published
Patent Application US 2004/0025 026, Filing Date: 2 August 2002.
[266] Karygiannis, T.， Eydt, B.， Barber, G.， Bunn, L.， and Phillips, T. (2007, April). Guidelines for
securing radio frequency identification (RFID) systems (NIST Special Publication 800-98).
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[267] Karygiannis, T.， and Owens, L. (2002, November). Wireless network security: 802.1 1，
Bluetooth and handheld devices (NIST Special Publication 800-48). Gaithersburg, MD, USA:
CSRC (Computer Security Resource Center), NIST.
[268] Kaufman, C, Perlman, R.， and Speciner, M. (1995). Network security: Private
communication in a public world. Upper Saddle River, New Jersey, USA: Prentice Hall, 39-
100.
[269] Keller, S. S. (2005, January 31). NIST- recommended random number generator based on
ANSI X9.31 Appendix A.2.4 using the 3 -key triple DES and AES algorithms, [Online].
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST. Available:
http://csrc.nist.gov/groups/STM/cavp/documents/rng/931rngext.pdf [2008, May 19].
[270] Kelsey, J.， Schneier, B.， Hall, C.， and Wagner, D. (1997， September 17-19). Secure
applications of low- entropy keys. Proceedings of the International Workshop on Information
Security (ISW '97j， LNCS 1396， Tats 皿 okuchi, Ishikawa, Japan, 121-134.
[271] Khatri, N.， and Tsang, E. W. K. (2003, April). Antecedents and consequences of cronyism in
organizations. Journal of Business Ethics, 43(4), 289-303.
[272] Khaw, L. T. (2001). Copyright law in Malaysia (2nd ed.). Kuala Lumpur, Malaysia: Malayan
Law Journal.
288
[273] Kini, A., and Choobineh, J. (1998, January 6-9). Trust in electronic commerce: Definition
and theoretical considerations. Proceedings of the 31st Hawaii International Conference on
System Sciences 1998 (HICSS 1998)， Kohala Coast, HI, USA, Vol. 4， 51-61.
[274] Kissel, R. (Ed.). (2006, April 25). Glossary of key information security terms (NIST IR 7298).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[275] Klein, D. V. (1990， August 27). "Foiling the cracker": A survey of, and improvements to,
password security (revised paper). Proceedings of the USENIX 2nd Security Workshop
Program, Portland, Oregon, USA, 5-14.
[276] Klimek, P., Hanel, R.， and Thurner, S. (2008, April 14). To how many politicians should
government be left?, [Online]. Cornell University Library, Ithaca, NY, USA. Available:
http://arxiv.org/pdf/0804.2202 [2008, July 18].
[277] Koblitz, N. (1994， September 2). A course in number theory and cryptography (2nd ed.)-
New York, NY, USA: Springer- Verlag New York, Inc.
[278] Komyunitipatentorebyu [コミ ュニテ イノ、 ァ ン 卜 レビュー] - Community patent review,
[Online]. (No date). JPO (Japan Patent Office). Available: http://www.cprtrial-iip.org [2008,
August 21].
[279] Konrad, K.， Fuchs, G., and Barthel, J. (1999， October 19-22). Trust and electronic commerce
- More than a technical problem. Proceedings of the 18th IEEE Symposium on Reliable
Distributed Systems 1999 (RELDIS 1999), Lausanne, Switzerland, 360-365.
[280] Kotadia, M. (2004, February 25). Gates predicts death of the password, [Online]. CNET
Networks, Inc. Available: http://news.cnet.com/Gates-predicts-death-of-the-password/2100-
1029_3-5 164733.html?tag=nw.4 [2008, August 23].
[281] Kormann, D. P., and Rubin, A. D. (2000, June). Risks of the Passport single signon protocol.
Computer Networks, 33{\-6), 51-58.
[282] Kucera, H.， and Francis, W. N. (1967). Computational analysis of present-day American
English. Providence, Rhode Island, USA: Brown University Press.
[283] Kuhn, D. R.， Hu, V. C， Polk, W. T.， and Chang, S.-J. (2001, February 26). Introduction to
public key technology and the federal PKI infrastructure (NIST Special Publication 800-32).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
289
[284] Kurokawa, K. (1988， July). Quality and innovation [Japanese electronics industry]. IEEE
Circuits and Devices Magazine, 4(4), 3-8.
[285] Kurokawa, K. (1990， May 13-18). Quality and innovation. Proceedings of the 1990 IEEE
International Conference on Robotics and Automation (ROBOT 1990), Cincinnati, OH, USA,
Vol. 3， 2180-2184.
[286] Kurokawa, K. (1991， January). Quality and innovation. IEEE Control Systems Magazine,
77(1), 47-51.
[287] Kurokawa, K. (1997， April/May). Modeling human interactions. IEEE Potentials, 16(2), Part
2, 26-28.
[288] Kurzban, S. A. (1985, Fall/Winter). Easily remembered passphrases - A better approach.
ACM SIGSAC Review, 3(2-4), (SIGSAC: Special Interest Group on Security, Audit and
Control), 10-21.
[289] Lampe, K. (Ed.)- (No date). Definitions of organized crime, [Online]. Available:
http://www.organized-crime.de/OCDEFl.htm [2008, July 21].
[290] Lamport, L. (1983， July). The weak Byzantine Generals Problem. Journal of the ACM, 30(3),
668-676.
[291] Lamport, L.， Shostak, R.， and Pease, M. (1982, July). The Byzantine Generals Problem. ACM
Transactions on Programming Languages and Systems, 4(3), 382-401.
[292] LAN/MAN Standards Committee. (2005, December 9). Part 3: Carrier sense multiple
access with collision detection (CSMA/CD) access method and physical layer specifications,
[Online]. New York, NY, USA: IEEE Computer Society. Available: Access
http://standards.ieee.Org/getieee802/802.3.html (to get)
http://standards.ieee.org/getieee802/download/802.3-2005_section2.pdf [2008, January 24].
[293] Landy, D.， McCue, K.， and Aronson, E. (1969， June). Beyond Parkinson's law: III. The effect
of protractive and contractive distractions on the wasting of time on subsequent tasks.
Journal of Applied Psychology, 53(3), Part 1， 23o-z39.
[294] Lauer, H. C. (No date). Discussion on Ph.D. thesis proposals in computing science, [Online].
Available: http://homes.cerias.purdue.edu/~spaf/Archive/Lauer.html [2008, August 2].
290
[295] Layman, M. D.， and Potter, G. W. (1997). Organized crime. Upper Saddle River, New Jersey,
USA: Prentice-Hall, Inc.
[296] Le， A. V.， Matyas, S. M.， Johnson, D. B.， and Wilkins, J. D. (1993). A public key extension
to the common cryptographic architecture. IBM Systems Journal, 32(3), 461-485.
[297] Le Quere, P. (2004, March 30). High speed random number generation. USPTO Issued
Patent US67 14955, Filing Date: 13 August 2001, Issue Date: 30 March 2004.
[298] Lee, K. W. (2003, June 5). Artificial neural network based Byzantine Agreement Protocol.
Unpublished master's thesis, Multimedia University, Bukit Beruang, Melaka, Malaysia.
[299] Lee, K. W. (2005a, July 4-6). Semantic error occurrences in the multimedia communications.
Proceedings of IASTED 2005 International Conference on Education & Technology
(ICET2005), Calgary, Alberta, Canada, 227-231.
[300] Lee, K. W. (2005b, September 19). An nPST dual in-line package switch (DIUDIP switch)
Unking all the little actuators into a single actuator, [Electronic Component, Layout-Design
of Integrated Circuit]. Available: Lee, K. W.， Bukit Beruang, Melaka, Malaysia.
[301] Lee, K. W. (2006a, June 27-28). Byzantine agreement for electronic commerce. Proceedings
of the Regional Computer Science Postgraduate Conference 2006 (USM-ReCSPC06),
Penang, Malaysia, 114-119.
[302] Lee, K. W. (200ob, November 5). 2 -dimensional key input method version 1.1, [CD-ROM,
Computer Program]. Available: Lee, K. W., Bukit Beruang, Melaka, Malaysia.
[303] Lee, K. W. (2006c, November 29). An Ethernet cable RJ45 switch for securing data
communication and storage using conventional dual in-line package switch (DIL/DIP
switch ), [Electronic Component, Layout-Design of Integrated Circuit]. Available: Lee, K. W.,
Bukit Beruang, Melaka, Malaysia.
[304] Lee, K. W. (2007a, January). Account amount estimator of multihash key using key
strengthening based on time response, [CD-ROM, Computer Program]. Available: Lee, K.
W., Bukit Beruang, Melaka, Malaysia.
[305] Lee, K. W. (2007b, October 19). An (nPST + reverse mPST) dual in-line package switch
(DIL/DIP switch ) with two bigger actuators linking two groups of switches activated
291
oppositely, [Electronic Component, Layout-Design of Integrated Circuit]. Available: Lee, K.
W., Bukit Beruang, Melaka, Malaysia.
[306] Lee, K. W. (2007c, December 11). An Ethernet cable RJ45 switch for securing data
communication and storage using Gray hill 78H02T (double 4PST DIP switch ), [Electronic
Component, Layout-Design of Integrated Circuit]. Available: Lee, K. W., Bukit Beruang,
Melaka, Malaysia.
[307] Lee, K. W. (2007d, December 11). An Ethernet cable RJ45 switch for securing data
communication and storage using Gray hill 78H01T (single 4PST DIP switch ), [Electronic
Component, Layout-Design of Integrated Circuit]. Available: Lee, K. W.， Bukit Beruang,
Melaka, Malaysia.
[308] Lee, K. W. (2007e, December 11). An Ethernet cable RJ45 switch for securing data
communication and storage using Gray hill 78F01T (single 2PST DIP switch ), [Electronic
Component, Layout-Design of Integrated Circuit]. Available: Lee, K. W.， Bukit Beruang,
Melaka, Malaysia.
[309] Lee, K. W. (2008a, May 8). An improved dual in-line (DIL) switch for securing data
communication and storage. Malaysia Patent Application PI 20070733, MylPO, Filing Date:
11 May 2007.
[310] Lee, K. W. (2008b, May 8). An improved dual in-line (DIL) switch for securing data
communication and storage. PCT Patent Application PCT/MY2008/000040, WIPO, Filing
Date: 8 May 2008.
[311] Lee, K. W. [李國 華]. (2008c, May 9). 應用於 保護 資料 通 訊 及存儲 安全 之 改良 式 雙行開
M [An improved dual in-line (DIL) switch for securing data communication and storage].
Taiwan (ROC) Patent Application TW097 117364， TIPO, Filing Date: 9 May 2008. (in
Chinese language).
[312] Lee, K. W. (2008d, June 2). A telephone cable RJ11 switch for securing data communication
and storage using Gray hill 78H01T (single 4PST DIP switch ), [Electronic Component,
Layout-Design of Integrated Circuit]. Available: Lee, K. W., Bukit Beruang, Melaka,
Malaysia.
[313] Lee, K. W. (2008e, June 2). A telephone cable RJ11 switch for securing data communication
and storage using Gray hill 78F01T (single 2PST DIP switch j, [Electronic Component,
292
Layout-Design of Integrated Circuit]. Available: Lee, K. W.， Bukit Beruang, Melaka,
Malaysia.
[314] Lee, K. W. (2008f, June 2). A USB switch for securing data communication and storage
using Gray hill 78H01T (single 4PST DIP switch ), [Electronic Component, Layout-Design of
Integrated Circuit]. Available: Lee, K. W.， Bukit Beruang, Melaka, Malaysia.
[315] Lee, K. W. (2008g, June 2). A USB switch for securing data communication and storage
using Gray hill 78F0 IT (single 2PST DIP switch), [Electronic Component, Layout-Design of
Integrated Circuit]. Available: Lee, K. W.， Bukit Beruang, Melaka, Malaysia.
[316] Lee, K. W. (2008h, July 25). Methods and systems to create big memorizable secrets and
their applications in information engineering. Malaysia Patent Application PI 20082771,
MylPO, Filing Date: 25 July 2008.
[317] Lee, K. W. (2008i, September 21). 2 -dimensional key input method with multihash key
version 2.0， [CD-ROM, Computer Program]. Available: Lee, K. W., Bukit Beruang, Melaka,
Malaysia.
[318] Lee, K. W. (2008j, September 21). Mobile ECC version 2.0， [CD-ROM, Computer Program].
Available: Lee, K. W., Bukit Beruang, Melaka, Malaysia.
[319] Lee, K. W. (2008k, September 21). 2 -factor multimedia key with multihash key version 1.0,
[CD-ROM, Computer Program]. Available: Lee, K. W.， Bukit Beruang, Melaka, Malaysia.
[320] Lee, K. W.， and Ewe, H. T. (2001, October 2001). Artificial neural networks based algorithm
for Byzantine Generals Problem. Proceedings of the MMU International Symposium on
Information and Communications Technologies 2001 ( MMU -Ml U SIC 2001), Petaling Jaya,
Selangor, Malaysia, pp. TS 1.4(1-4).
[321] Lee, K. W., and Ewe, H. T. (2002, July 17-19). Artificial neural network based Byzantine
Agreement Protocol. Proceedings of the 6th IASTED International Conference on Artificial
Intelligence and Soft Computing (IASTED-ASC 2002 J, Banff, Alberta, Canada, 368-373.
[322] Lee, K. W., and Ewe, H. T. (2003， October 2-3). Faulty node detection in the tripartite ANN
based BAP. Proceedings of the MMU International Symposium on Information and
Communications Technologies 2003 (MMU-M2USIC 2003 ), Petaling Jaya, Selangor,
Malaysia, 45-48.
293
[323] Lee, K. W.， and Ewe, H. T. (2006, November 3-6). Coinware for multilingual passphrase
generation and its application for Chinese language password. Proceedings of the 2006
International Conference on Computational Intelligence and Security (CIS 2006)，
Guangzhou, Guangdong, China, 1511-1514 (Part 2).
[324] Lee, K. W., and Ewe, H. T. (2007a, August). Multiple hashes of single key with passcode for
multiple accounts. Journal ofZhejiang University Science A (JZUS-A), 8(8), 1183-1190.
[325] Lee, K. W.， and Ewe, H. T. (2007b, November 1). Performance study of Byzantine
Agreement Protocol with artificial neural network. Information Sciences, 777(21), 4785-4798.
[326] Lee, K. W., and Ewe, H. T. (2007c, November 12-15). Speeding up the Byzantine
Agreement Protocol with artificial neural network (BAP-ANN) using modified
backpropagation learning algorithm (BPLA). Proceedings of the Malaysia-Japan
International Symposium on Advanced Technology 2007 (M J IS AT 2007), Kuala Lumpur,
Malaysia, pp. 65 & 169 [CD-ROM, Technical Session T5-5: AI Applications III].
[327] Lee, K. W., and Ewe, H. T. (2007d, December 15-19). Passphrase with semantic noises and a
proof on its higher information rate. Proceedings of the International Conference on
Computational Intelligence and Security Workshops (CISW 2007), Harbm, Heilon^jiang,
China, 652-655.
[328] Lee, K. W.， Radhakrishna, G.， and Khaw, L. T. (2007, November 19-20). The proof of
copyright ownership using digital timestamp in Malaysia. Proceedings of the MMU
International Symposium on Information and Communications Technologies 2007 (MMU-
M2USIC 2007)， Petaling Jaya, Selangor, Malaysia, pp. 20 & 37， [CD-ROM, file (f.) TS3A-
l.pdf]. Multimedia University, Cyberjaya (Selangor) & Bukit Beruang (Melaka), Malaysia.
[329] Lee, K. W.， and Tan, A. W. C. (2006a, November 19). Multilingual key input method version
1.0, [CD-ROM, Computer Program]. Available: Lee, K. W., Bukit Beruang, Melaka,
Malaysia.
[330] Lee, K. W., and Tan, A. W. C. (2006b, November 24). MobileECC version 1.2, [CD-ROM,
Computer Program]. Available: Lee, K. W., Bukit Beruang, Melaka, Malaysia.
[331] Lee, K. W., Teh, C. E.， and Tan, Y. L. (2006, May 30-31). Decrypting English text using
enhanced frequency analysis. Seminar conducted at the meeting of the National Seminar on
Science, Technology and Social Sciences (STSS 2006), Kuantan, Pahang, Malaysia.
294
[332] Lee, S.-C, Tang, D.， Chen, C.-T.， and Fang, S.-C. (2002, May). Finger-reading: Exploring
the information field. The International Journal of Healing and Caring, 2(2), 1-25.
[333] Lee, W. C. Y. (1995). Mobile cellular telecommunications: Analog and digital systems (2nd
ed.). Singapore: McGraw-Hill Book Company, 463-486.
[334] Lemley, M. A., and Moore, K. A. (2004). Ending abuse of patent continuations. Boston
University Law Review, 84， 63-118.
[335] Lemley, M. A., and Sampat, B. N. (2007, November 9-10). Is the patent office a rubber
stamp? Proceedings of the 2nd Annual Conference on Empirical Legal Studies (CELS 2007),
New York, NY, USA.
[336] Li, J.-R, Hu， G.-R, Wang, R.-H., and Dai, L.-R. (2005, March 18-23). Sliding window
smoothing for maximum entropy based intonational phrase prediction in Chinese.
Proceedings of the 2005 IEEE International Conference on Acoustics, Speech, and Signal
Processing (ICASSP V5), Philadelphia, PA, USA, 285-288.
[337] Liataud, J. P., and Moloney, M. L. (1983， March 8). DIP switch. USPTO Issued Patent
US4376234, Filing Date: 5 May 1981, Issue Date: 8 March 1983.
[338] Licensing revenue remains steady, [Online]. (No date). Baltimore, Maryland, USA: School
of Medicine, Johns Hopkins University. Available:
http://www.hopkinsmedicine.org/webnotes/licensing/0210.cfm [2008, September 3].
[339] Lilly, G. M. (2004, December 7). Device for and method of one-way cryptographic hashing.
USPTO Issued Patent US6829355, Filing Date: 5 March 2001, Issue Date: 7 December 2004.
[340] Limayem, M.， Khalifa, M.， and Chin, W. W. (2004, November). Factors motivating software
piracy: A longitudinal study. IEEE Transactions on Engineering Management, 57(4), 414-
425.
[341] Lin, H. C. (1999， October 19). Dual inline package switch. USPTO Issued Patent
US5967302, Filing Date: 6 March 1998， Issue Date: 19 October 1999.
[342] Liu, Q. Y. [刘 清彦] (Trans,), (2001, June). 補 圆曲线 公钢密 码导引 talents]. Taipei
[台北]， Taiwan (ROC) [中华 民国]: Lin Yu Cultural Enterprise Co. Ltd. [林 郁文 化 事业有
限 公司], (in Chinese language).
295
[343] Livingston, J. (1996). Crime & Criminology. Upper Saddle River, New Jersey, USA:
Prentice-Hall, Inc.
[344] Lockard, J. L. (1977， March 1^;. Miniature switch with substantial wiping action. USPTO
Issued Patent US4012608, Filing Date: 25 March 1975, Issue Date: 15 March 1977.
[345] Lockard, J. L. (1979， September 18). Impedance programming DIP switch assembly. USPTO
Issued Patent US4168404, Filing Date: 3 May 1978, Issue Date: 18 September 1979.
[346] Log This, [Online]. (No date). Available:
http://membersJycos.co.uk/wuul/logthis/readme.html [2008, March 25].
[347] Low, S. H.， and Maxemchuk, N. F. (1998， May). Performance comparison of two text
marking methods. IEEE Journal on Selected Areas in Communications, 7(5(4), 5ol-572.
[348] Lu， C.-S. (2002, December 9-11). Wireless multimedia error resilience via a data handling
technique. Proceedings of the 2002 IEEE Workshop on Multimedia Signal Processing, St.
Thomas, US Virgin Islands, USA, 316-319.
[349] Lu， C-S. (2005). Multimedia security: Steganography and digital watermarking techniques
for protection of intellectual property. Hershey, PA, USA: Idea Group Publishing.
[350] Luo, H.， and Henry, P. (2003, September 7-10). A common password method for protection
of multiple accounts. Proceedings of the 14th IEEE 2003 International Symposium on
Personal Indoor and Mobile Radio Communication (PIMRC 2003 j, Beijing, China, vol. 3，
2749-2754.
[351] Luo, H. Y. [罗华 炎]. (1990). 简明汉 语语法 [Concise Chinese grammar]. Cheras, Kuala
Lumpur, Malaysia: Yakin [雅 景]. (in Chinese language).
[352] Luo, H. Y. [罗华 炎]. (2003). 现代汉 语语法 [Modern Chinese grammar]. Ipoh, Perak,
Malaysia: Seni Hijau [艺 青], (in Chinese language).
[353] Lyons-Burke, K. (2000, October). Federal agency use of public key technology for digital
signatures and authentication (NIST Special Publication 800-25). Gaithersburg, MD, USA:
CSRC (Computer Security Resource Center), NIST.
296
[354] MacGregor, W.， Schwarzhoff, T.， and Mehta, K. (2007, June). A scheme for PIV visual card
topography (NIST Special Publication 800-104). Gaithersburg, MD， USA: CSRC (Computer
Security Resource Center), NIST.
[355] MacKenzie, P. D.， Oprea, A.， and Reiter, M. K. (2003, October 27-30). Automatic generation
of two-party computations (extended abstract). Proceedings of the 10th ACM Conference on
Computer and Communications Security 2003, Washington, D.C., USA, 210-219.
[356] MacKenzie, P. D.， and Reiter, M. K. (2001a, May). Networked cryptographic devices
resilient to capture (Tech. Rep. No. DIMACS TR 2001-19), [Online]. New Jersey, USA:
Rutgers (The State University of New Jersey), DIMACS (Center for Discrete Mathematics &
Theoretical Computer Science). Available:
http://dimacs.rutgers.edu/TechnicalReports/abstracts/2001/2001-19.html [2008, July 16].
[357] MacKenzie, P. D.， and Reiter, M. K. (2001b, May 14-16). Networked cryptographic devices
resilient to capture. Proceedings of the 2001 IEEE Symposium on Security and Privacy (S&P
2001), Oakland, CA， USA, 12-25.
[358] MacKenzie, P. D.， and Reiter, M. K. (2001c, August 19-23). Two-party generation of DSA
signatures (extended abstract). Proceedings of the 21st Annual International Crypto logy
Conference on Advances in Cryptology (CRYPTO 2001), LNCS 2139, Santa Barbara, CA，
USA, 137-154.
[359] MacKenzie, P. D.， and Reiter, M. K. (2002, June 26). Methods and apparatus for two-party
generation of DSA signatures. US Published Patent Application 2003/0059041, Filing Date:
26 June 2002.
[360] MacKenzie, P. D.， and Reiter, M. K. (2003, November). Networked cryptographic devices
resilient to capture. International Journal of Information Security, 2(1), 1-20.
[361] MacKenzie, P. D.， and Reiter, M. K. (2004, August). Two-party generation of DSA
signatures. International Journal of Information Security, 2(3-4), 218-239.
[362] MacKenzie, P. D.， and Reiter, M. K. (2006, December 12). Methods and apparatus for
providing networked cryptographic devices resilient to capture. USPTO Issued Patent
US7149311, Filing Date: 7 February 2002, Issue Date: 12 December 2006.
297
[363] Macuch, P. L. (2005, November 22). Coaxial and DSL cable switch for controlling a
computer connection to the Internet. USPTO Issued Design Patent D511749, Filing Date: 17
November 2003, Issue Date: 22 November 2005.
[364] Maghiros, L， Punie, Y., Delaitre, S.， Lignos, E.， Rodriguez, C, Ulbrich, M., Cabrera, M.，
Clements, B.， Beslay, L.， and van Bavel, R. (2005). Biometrics at the frontiers: Assessing the
impact on society (Report No. EUR 21585 EN). Seville, Se villa, Spain: European
Commission, Joint Research Centre (JRC), Institute for Prospective Technological Studies
(IPTS).
[365] Malackowski, J. E.， and Barney, J. A. (2008, June). What is patent quality? A merchant
banc's perspective, les Nouvelles, [Online] 2008Q\mo), 123-134. Licensing Executives
Society International (LES I). Available:
http://www.lesi.org/content/article_of_the_month.aspx [2008, August 18].
[366] Maltoni, D.， Maio, D.， Jain, A. K.， and Prabhakar, S. (2003). Handbook of fingerprint
recognition. New York, NY, USA: Springer- Verlag New York, Inc.
[367] Mann an, M.， and van Oorschot, P. C. (2007, September 18-21). Security and usability: The
gap in real-world online banking. Proceedings of the New Security Paradigms Workshop
2007 (NSPW'07), North Conway, New Hampshire, USA, 1-14.
[368] Mann an, M.， and van Oorschot, P. C. (2008, July 29). Digital objects as passwords.
Proceedings of the 3rd US EN IX Workshop on Hot Topics in Security (HotSec '08), San Jose,
California, USA, 1-6.
[369] Manoharan, S.， and Wu, J. (2007, January 2-7). Software licensing: A classification and case
study. Proceedings of the 1st International Conference on the Digital Society 2007 (I CDS
2007), Le Gosier, Guadeloupe, French Caribbean, paper 33.
[370] Mambo, M.， Usuda, K.， and Okamoto, E. (1996, September 20). Proxy signature: Delegation
of the power to sign messages. IEICE Transactions on Fundamentals of Electronics,
Communications and Computer Sciences, E79-A(9), 1338-1353.
[371] Manber, U. (1996). A simple scheme to make passwords based on one-way functions much
harder to crack. Computers & Security, 75(2), 171-176.
298
[372]
Markoff, J. (2008, October 20). A robot network seeks to enlist your computer. The New
York Times, [Online]. Available:
http://www.nytimes.com/2008/10/21/technology/internet/21botnet.html [2008, October 23].
[373] Marshall, D. (2003). .NET security programming. Indianapolis, IN, USA: Wiley.
[374] Matias, Y.， Mayer, A., and Silberschatz, A. (1997， December 8-11). Lightweight security
primitives for e-commerce. Proceedings of the USENIX Symposium on Internet Technologies
and Systems 1997, Monterey, California, USA, 95-102.
[375] Matthews, P. H. (1997). The concise Oxford dictionary of linguistics. New York, NY:
Oxford University Press.
[376] Maurer, U. (2004, June). New approaches to digital evidence. Proceedings of the IEEE,
92(6)， 933-947.
[377] Maxim, P. S.， and Whitehead, P. C. (1998). Exploring crime (4th ed.). Woburn, MA, USA:
Butterworth-Heinemann.
[378] McCallister, E.， and Ferraiolo, H.. (2006, August). Personal identity verification
demonstration summary (NIST IR 7337). Gaithersburg, MD， USA: CSRC (Computer
Security Resource Center), NIST.
[379] McClure, S.， Scambray, J.， and Kurtz, G. (2001, September 26). Hacking exposed: Network
security secrets & solutions (3rd ed.)- Berkeley, California, USA: Osborne / McGraw-Hill.
[380] McCulligh, M. R. (2003, November 4). Password generation method and system. USPTO
Issued Patent US6643784, Filing Date: 14 December 1998, Issue Date: 4 November 2003.
[381] MCMC. (No date). Register of recognised date/time stamp services (Section 70). In Registers
Under The Digital Signature Act 1997, [Online]. Malaysian Communications and
Multimedia Commission (MCMC). Available:
http://www.mcmc.gov.my/registers/DigitalSignature/RecognizesDateTimeStampService/ind
ex. asp [2007, May 23].
[382] McMillan, R. (2008a, June 10). Microsoft hires anti-phishing crusader, [Online]. CIO.com.
Available: http://www.cio.com/article/print/390413 [2008, September 25].
299
[383] McMillan, R. (2008b, July 1). Trojan lurks, waiting to steal admin passwords, [Online].
CIO.com. Available: http://www.cio.com/article/print/421463 [2008, September 10].
[384] McNamara, J. (2003). Secrets of computer espionage: Tactics and countermeasures.
Indianapolis, Indiana, USA: Wiley Publishing, Inc.
[385] MDC. (2002a, September). Copyright Act and Regulations (Act 332). In Laws relating to
intellectual property and cy be daws in Malaysia (pp. 288-408). Kuala Lumpur, Malaysia:
MDC.
[386] MDC. (2002b, September). Digital Signature Act and Regulations (Act 562). In Laws
relating to intellectual property and cy be daws in Malaysia (pp. 479-591). Kuala Lumpur,
Malaysia: MDC.
[387] MDC. (2002c, September). Industrial Designs Act and Regulations (Act 552). In Laws
relating to intellectual property and cy be daws in Malaysia (pp. 409-478). Kuala Lumpur,
Malaysia: MDC.
[388] MDC. (2002d, September). Layout-Designs of Integrated Circuits Act and Regulations (Act
6001). In Laws relating to intellectual property and cy be daws in Malaysia (pp. 606-625).
Kuala Lumpur, Malaysia: MDC.
[389] MDC. (2002e, September). Patents Act and Regulations (Act 291). In Laws relating to
intellectual property and cy be daws in Malaysia (pp. 157-287). Kuala Lumpur, Malaysia:
MDC.
[390] Menezes, A. J., Oorschot, P. C. V., and Van stone, S. A. (1996). Handbook of applied
cryptography. Boca Raton, FL, USA: CRC Press.
[391] Meurer, M. J., and Bessen, J. E. (2008, March 7). Do patents perform like property?,
[Online]. Boston University School of Law Working Paper No. 08-08. Available:
http://papers.ssrn.com/sol3/papers.cfm ?abstrac ᄂ id=l 103143 [2008, August 20].
[392] Meyer, F. J., and Pradhan, D. K. (1991， April). Consensus with dual failure modes. IEEE
Transactions on Parallel and Distributed Systems, 2(2), 214 - 222.
[393] Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our
capacity for processing information. The Psychological Review, 63(2), 81-97.
300
[394] Mitchell, E. D. (1974). Psychic exploration: A challenge for science. New York, NY, USA:
G. P. Putnam's Sons.
[395] Mittelholzer, T. (1999， September 29 - October 1). An information-theoretic approach to
steganography and watermarking. Proceedings of the 3rd International Workshop on
Information Hiding (IH '99), LNCS 1768, Dresden, Saxony, Germany, 1-16.
[396] Mohanty, S. P. (1999). Digital watermarking: A tutorial review, [Online, Tech. Rep.].
Bangalore, Bayaluseeme, Karnataka, India: Indian Institute of Science. Available:
http://citeseer.ist.psu.edu/572262.html [2008, July 17].
[397] Mollin, R. A. (2007a). Codes: The guide to secrecy from ancient to modem times. Boca
Raton, FL， USA: Chapman & Hall/CRC， Taylor & Francis Group.
[398] Mollin, R. A. (2007b). An introduction to cryptography (2nd ed.). Boca Raton, FL， USA:
Taylor & Francis Group, Chapman & Hall/CRC.
[399] Moseley, B. £. (2006, February 2). Method and system for generating passwords. USPTO
Published Patent Application US 2006/0026439, Filing Date: 2 August 2004.
[400] Mossinghoff, G. J. (2005, July 20；. Testimony in the US Senate - Perspectives on patents:
Harmonization and other matters, [Online]. Committee on the Judiciary of the United States
Senate. Available: http://judiciary.senate.gov/testimony.cfm?id=1582&wi ᄂ id=4547 [2007,
October 22].
[401] Mother tongue/bilingual literacy programme for ethnic minorities. (No date). UNESCO
Bangkok, [Online]. Available: http://www.unescobkk.org/index. php?id=222 [2008, October
4].
[402] Moulin, P., and O' Sullivan, J. A. (2003, March). Information-theoretic analysis of
information hiding. IEEE Transactions on Information Theory, 49(3), 563-593.
[403] MylPO. (No date). Copyright, [Online]. Intellectual Property Corporation of Malaysia
(MylPO). Available:
http://www.mipc.gov.my/index.php?option=com_content&task=view&id=18&itemid=19
[2007, May 20].
[404] Nash, A.， Duane, W.， Joseph, C.， and Brink, D. (2001). PKI: Implementing and managing e-
security. Berkeley, CA， USA: Osborne/McGraw-Hill.
301
[405] Nasheri, H. (2004, December 6). Economic espionage and industrial spying. Cambridge,
Cambridgeshire, UK: Cambridge University Press.
[406] Nechvatal, J. (1991, April). Public-key cryptography (NIST Special Publication 800-2).
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[407] Nguyen, D.， and Widrow, B. (1989， June 18-22). The track backer-upper: An example of
self-learning in neural networks. Proceedings of the International Joint Conference on
Neural Networks 1989 (IJCNN 1989), Washington, D.C., USA, Vol. 2， 357-363.
[408] Nguyen, D.， and Widrow, B. (1990, June 17-21). Improving the learning speed of two-layer
neural networks by choosing initial values of the adaptive weights. Proceedings of the
International Joint Conference on Neural Networks 1990 (IJCNN 1990), San Diego, CA,
USA, Vol. 3, 21-26.
[409] Nichols, R. K. (1999). ICSA guide to cryptography. Blacklick, OH, USA: McGraw-Hill.
[410] NIST. (1985a, May 30). Computer data authentication (NIST FIPS Pub 113). Gaithersburg,
MD, USA: CSRC (Computer Security Resource Center), NIST.
[411] NIST. (1985b, May 30). Password usage (FIPS Pub 112). Gaithersburg, MD, USA: CSRC
(Computer Security Resource Center), NIST.
[412] NIST. (1992, April 27). Key management using ANSI X9.17 (NIST FIPS Pub 171).
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[413] NIST. (1993, October 5). Automated password generator (NIST FIPS Pub 181).
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[414] NIST. (1994a, February 9). Escrowed encryption standard (NIST FIPS Pub 185).
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[415] NIST. (1994b, September 28). Guidelines for the use of advanced authentication technology
alternatives (NIST FIPS Pub 190). Gaithersburg, MD, USA: CSRC (Computer Security
Resource Center), NIST.
[416] NIST. (1995a, April 17). Secure hash standard (NIST FIPS Pub 180-1). Gaithersburg, MD,
USA: CSRC (Computer Security Resource Center), NIST.
302
[417] NIST. (1995b, October). An introduction to computer security: The NIST handbook (NIST
Special Publication 800-12). Gaithersburg, MD, USA: CSRC (Computer Security Resource
Center), NIST.
[418] NIST. (1997， February 18). Entity authentication using public key cryptography (NIST FIPS
Pub 196). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[419] NIST. (2000, January 27). Secure signature standard (DSS) (NIST FIPS Pub 186-2 (+
Change Notice)). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center),
NIST.
[420] NIST. (2001, May 2^). Security requirements for cryptographic modules (NIST FIPS Pub
140-2). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[421] NIST. (2002a, March 6). The keyed-hash message authentication code (HMAC) (NIST FIPS
Pub 198). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[422] NIST. (2002b, August 1). Secure hash standard (NIST FIPS Pub 180-2 (+ Change Notice to
include SHA-224)). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center),
NIST.
[423] NIST. (2004, February). Standards for security categorization of federal information and
information systems (NIST FIPS Pub 199). Gaithersburg, MD, USA: CSRC (Computer
Security Resource Center), NIST.
[424] NIST. (2005a, July). Questions and answers about the certification and accreditation of PIV
card issuing organizations, [Online]. Gaithersburg, MD, USA: CSRC (Computer Security
Resource Center), NIST. Available: http://csrc.nist.gov/publications/nistpubs/800-79/sp800-
79Q-As.pdf [2008, May 19].
[425] NIST. (2005b, July). Questions and answers regarding certification and accreditation of PIV
card issuing organizations, [Online]. Gaithersburg, MD, USA: CSRC (Computer Security
Resource Center), NIST. Available: http://csrc.nist.gov/publications/nistpubs/800-79/sp800-
79Q-As-Part2.pdf [2008, May 19].
[426] NIST. (2006a, March). Minimum secuirty requirements for federal information and
information systems (NIST FIPS Pub 200). Gaithersburg, MD, USA: CSRC (Computer
Security Resource Center), NIST.
303
[427] NIST. (2006b, March). Personal identity verification (PIV) of federal employees and
contractors (NIST FIPS Pub 201-1 (Change Notice 1)). Gaithersburg, MD， USA: CSRC
(Computer Security Resource Center), NIST.
[428] NIST. (2006c, March 13). Secure signature standard (DSS) (draft) (NIST FIPS Pub 186-3
(Draft)). Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[429] NIST. (2006d, June). Roadmap to NIST information security documents, [Online Brochure].
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST. Available:
http://csrc.nist.gov/publications/CSD_DocsGuide_Trifold.pdf [2008, May 19].
[430] NIST. (2007a, March). Guide to NIST information security documents, [Online].
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST. Available:
http://csrc.nist.gov/publications/CSD_DocsGuide.pdf [2008, May 19].
[431] NIST. (2007b, June 12). Secure hash standard (SHS) (draft) (NIST FIPS Pub 180-3 (Draft)).
Gaithersburg, MD, USA: CSRC (Computer Security Resource Center), NIST.
[432] NIST. (2007c, June 12). The keyed-hash message authentication code (腿 AC) (draft) (NIST
FIPS Pub 198-1 (Draft)). Gaithersburg, MD, USA: CSRC (Computer Security Resource
Center), NIST.
[433] NIST. (2007d, July 13). Security requirements for cryptographic modules (draft ) (NIST
FIPS Pub 140-3 (Draft)). Gaithersburg, MD, USA: CSRC (Computer Security Resource
Center), NIST.
[434] NIST. (2008, February 7). Implementation guidance for FIPS PUB 140-2 and the
cryptographic module validation program (NIST FIPS Pub 140-2 IG). Gaithersburg, MD,
USA: CSRC (Computer Security Resource Center), NIST.
[435] Noble, K. A. (1994). Changing doctoral degrees: An international perspective. Buckingham,
Buckinghamshire, UK: SRHE (The Society for Research into Higher Education) & Open
University Press.
[436] Ocean Tomo, [Online]. (No date). Available: http://www.oceantomo.com [2008, August 21].
[437] Ogletree, T. W. (2000). Practical firewalls. Indianapolis, Indiana, USA: Macmillan
Computer Publishing, Que Corporation.
304
[438] Oppliger, R.， and Rytz, R. (2003, September-October). Digital evidence: Dream and reality.
IEEE Security & Privacy Magazine, 7(5), 44-48.
[439] Oram, A. (2008, February). Peer to Patent needs your expertise. Communications of the ACM,
57(2)， 19-20.
[440] Ostrander, S.， and Schroeder, L. (1974). Handbook of psychic discoveries. New York, NY,
USA: Berkley Publishing Corporation.
[441] Pankaj, S. (2005). Hacking. Darya Ganj, New Delhi, India: APH Publishing Corporation.
[442] Paradowski, M. B. (No date). The benefits of multilingualism, [Online]. Bilingual/Bicultural
Family Network. Available: http://www.biculturalfamily.org/benefitsofmultilingualism.html
[2008, October 4].
[443] Parapsychological Association, [Online]. (No date). Available: http://www.parapsych.org
[2008, October 25].
[444] Parkinson, C. N. (1958). Parkinson's Law: Or the Pursuit of Progress, [Online]. Available:
http://www.adstockwebxom/business-lore/Parkinson's_Law.htm [2008, July 18].
[445] Parkinson, C. N. (2002, September 5). Parkinson's Law: Or the pursuit of progress (new ed.).
London, UK: Penguin Books Ltd.
[446] Parnas, D. L. (2007, November). Stop the numbers game. Communications of the ACM,
5 り (11)， 19-21.
[447] PasswordResearch.com, [Online]. (No date). Available: http://www.passwordresearch.com
[2008, September 10].
[448] Patel, D.， and Luo, X. (2007, September 28-29). Take a close look at phishing. ACM
Proceedings of the 4th Annual Conference on Information Security Curriculum Development
2007, Kennesaw, Georgia, USA, 210-213.
[449] Pathak, V., and Iftode, L. (2006, March). Byzantine fault tolerant public key authentication in
peer-to-peer systems. Computer Networks: The International Journal of Computer and
Telecommunications Networking, 50(4), 579-596.
305
[450] Pease, M.， Shostak, R.， and Lamport, L. (1980, April). Reaching agreement in the presence
of faults. Journal of the ACM, 27(2), 228-234.
[451] Peer to Patent - Community patent review, [Online]. (No date). Available:
http://www.peertopatent.org [2008, August 21].
[452] Pelzl, J.， Wollinger, T.， and Paar, C. (2004, April 5-7). High performance arithmetic for
special hyperelliptic curve cryptosystems of genus two. Proceedings of the International
Conference on Information Technology: Coding and Computing 2004 (ITCC 2004), Las
Vegas, Nevada, USA, Vol. 2， 513-517.
[453] Penguin Popular Classics. (1997). Arabian nights: AM Baba and the forty thieves. London,
UK: Penguin Books, 100-126.
[454] Petitcolas, F. A. P., Anderson, R. J.， and Kuhn, M. G. (1999， July). Information hiding - A
survey. Proceedings of the IEEE: Special Issue on Protection of Multimedia Content, 87(1),
1062-1078.
[455] Pfitzmann, B. (1996). Digital signature schemes: General framework and fail-stop
signatures. Berlin, Germany: Springer- Verlag.
[456] Pfitzmann, B.， and Waidner, M. (1990). Formal aspects of fail-stop signatures (Tech. Rep.
No. Interner Bericht 22/90), [Online]. Karlsruhe, Baden-Wiirttemberg, Germany: Universitat
Karlsruhe, Fakultat fiir Informatik. Available: http://ftp.arnes.si/pub/packages/crypto-
papers/PfWa_90FSS_formal_IB.ps.gz [2008， July 17].
[457] PGP Corporation. (2006). PGP Desktop 9.0 for windows user's guide. Palo Alto, California,
USA: PGP Corporation, 229-232.
[458] Phillips, E. M.， and Pugh, D. S. (2005, June 1). How to get a PhD (4th ed.). Maidenhead,
Berkshire, UK: McGraw-Hill Education, Open University Press.
[459] Pierer, H. V., and Oetinger, B. V. (Eds.)- (2002). A passion for ideas: How innovators create
the new and shape our world. West Lafayette, Indiana, USA: Purdue University Press.
[460] Piatt, J. R. (2006a, August). Benefits of membership: Today's patents rely on IEEE-published
science. IEEE Microwave Magazine, 7(4)， 76-77.
306
[461] Piatt, J. R. (2006b, November). Patents - Today's patents rely on IEEE-published science.
IEEE Signal Processing Magazine, 23(6), 8 & 12.
[462] Polk, W. T.， Dodson, D. F., and Burr, W. E. (2007, August). Cryptographic algorithms and
key sizes for personal identity verification (NIST Special Publication 800-78-1).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[463] Popular Book [大 众书 局]. (2003). 汉语拼 音 学习卡 [Learning cards of Hanyu Pinyin].
Singapore: Popular Book, (in Chinese language).
[464] Pratt, W. (1997， March 5). Graduate school survival guide, [Online]. Available:
http://www.math.waikato.ac.nz/~seano/grad-school-advice.html [2008, July 21].
[465] Prepared testimony of Adam Jajfe for 15 Feb. 07. (2007, February 15). IPBiz.com. Available:
http://ipbiz.blogspot.com/2007/02/prepared-testimony-of-adam-jaffe-for-15.html;
http://judiciary.house.gov/media/pdfs/jaffe070215.pdf [2008, August 20].
[466] Project 10 wo , [Online]. (2008). Google. Available:
http://www.projectlOtothelOO.com/how_it_works.html [2008, October 13].
[467] Raina, K. (2003). PKI security solutions for the enterprise: Solving HIPAA, E-Paper Act, and
other compliance issues. Indianapolis, Indiana, USA: Wiley Publishing, Inc.
[468] Re, V., Borean, C， Bozzi, C， Carassiti, V.， Ramusino, A. C， Piemontese, L.， Breon, A. B.，
Brown, D.， Clark, A. R.， Goozen, R， Hernikl, C, Kerth, L. T.， Gritsan, A., Lynch, G.，
Perazzo, A., Roe, N. A., Zizka, G.， Roberts, D.， Schieck, J., Brenna, E.， Citterio, A., Lanni, F.，
Palombo, R， Ratti, L.， Manfredi, P. R， Angelini, C， Batignani, G.， Bettarini, S.， Bondioli, M.，
Bosi, R， Bucci, R， Calderini, G.， Carpinelli, A., Ceccanti, M.， Forti, R， Gagliardi, D.， Giorgi,
M. A., Lusiani, A., Mammini, P., Morganti, M.， Morsani, F.， Neri, N.， Paoloni, E.， Profeti, A.,
Rama, M.， Rizzo, G., Sandrelli, F.， Simi, G., Triggiani, G., Walsh, J.， Burchat, P., Cheng, C.，
Kirkby, D.， Meyer, T. I.， Roat, C， Bona, A.， Bianchi, F.， Gamba, D.， Trapani, P., Bosisio, L.，
Delia Ricca, G.， Dittongo, S.， Lanceri, L.， Pompili, A.， Poropat, P., Rashevskaia, L， Vuagnin,
G., Burke, S., Callahan, D.， Campagnari, C， Dahmes, B.， Hale, D.， Hart, P., Kuznetsova, N.，
Kyre, S., Levy, S.， Long, O., May, J.， Mazur, M.， Richman, J.， Verkerke, W.， Witherell, M.，
Beringer, J.， Eisner, A. A., Frey, A.， Grillo, A. A.， Grothe, A.， Johnson, R. P., Kroeger, W.,
Lockman, W. S.， Pulliam, T.， Rowe, W.， Schmitz, R. E.， Seiden, A.， Spencer, E. N.， Turri,
M., Walkoviak, W., Wilder, M.， Wilson, M.， Charles, E.， Elmer, P., Nielsen, J.， Orejudos, W.，
Scott, L， Zobernig, H., and Laplace, S. (2002, December). The BaBar silicon-vertex tracker:
307
performance, running experience, and radiation-damage studies. IEEE Transactions on
Nuclear Science, 49(6)， Part 2, 3284-3289.
[469] Ren, J., and Taylor, R. N. (2007, June). Automatic and versatile publications ranking for
research institutions and scholars. Communications of the ACM, 50(6), 81-85.
[470] Research Evaluation, [Online]. (No date). Available: http://www.scipol.co.uk/re.htm [2008,
September 3].
[471] Risen, W. M.， and Covello, D. F. (2000, January 25). Method of protecting against a change
in value of intellectual property, and product providing such protection. USPTO Issued
Patent US6018714, Filing Date: 8 November 1997， Issue Date: 25 January 2000.
[472] Ritter, T. (2001, June 27). The redundancy of English, [Online]. Available:
http://www.ciphersbyritter.com/NEWS6/REDUN.HTM [2008, July 21].
[473] Rivest, R. (1992, April). Request for comments (1321): The MD5 Message-Digest Algorithm
(RFC 1321). Sterling, Virginia, USA: Network Working Group, The Internet Engineering
Task Force (IETF).
[474] Rivest, R. L.， Shamir, A.， and Adleman, L. (1978, February). A method for obtaining digital
signatures and public key cryptosystems. Communications of the ACM, 27(2), 120-126.
[475] Rivest, R. L.， Shamir, A., and Adleman, L. M. (1983， September 20). Cryptographic
communications system and method. USPTO Issued Patent US4405829, Filing Date: 14
December 1977， Issue Date: 20 September 1983.
[476] Roellgen, C. B. (No date). Scalable polymorphic hash function, [Online]. Available:
http://www.pmc-ciphers.com [2008, July 16].
[477] Roellgen, B. (2005, June 15). Wdhrend der laufzeit verdnderbare kryptographische methode
[Cryptographic method modifiable during run time]. EPO Issued Patent EP1069508, Filing
Date: 4 July 2000, Issue Date: 15 June 2005. (Original work publish 2005 in German
language).
[478] Ross, B.， Jackson, C.， Miyake, N.， Boneh, D.， and Mitchell, J. C. (2005, July 31 - August 5).
Stronger password authentication using browser extensions. Proceedings of the 14th U SEN IX
Security Symposium (SEC 2005), Baltimore, MD, USA, 17-32.
308
[479] Ruan, F. M, [阮方 民]， and Wang, X. [王 晓]. (2005, August). 有组织 犯罪 理论- 中国 黑社
会 性 质组织 犯 非 防 治 研' M [New theorv of organized crime - The research on prevention of
ganglands' crime in China]. Hangzhou [杭 州]， Zhejiang Province [浙江 省]， China [中国]:
Zhejiang University Press [浙江 大学 出 社], (in Chinese language).
[480] Rubin, K.， and Silverberg, A. (2003, August 17-21). Torus-based cryptography. Proceedings
of the 23rd Annual International Cryptology Conference 2003 (CRYPTO 2003), LNCS 2729,
Santa Barbara, California, USA, 349-365.
[481] Rugg, G.， and Petre, M. (2004). The unwritten rules of PhD research. Maidenhead,
Berkshire, UK: McGraw-Hill Education, Open University Press.
[482] Rukhin, A.， Soto, J.， Nechvatal, J.， Smid, M.， Barker, E.， Leigh, S., Levenson, M.， Vangel,
M.， Banks, D.， Heckert, A.， Dray, J.， and Vo， S. (2001, May 15). A statistical test suite for
random and pseudorandom number generators for cryptographic applications (NIST Special
Publication 800-22). Gaithersburg, MD， USA: CSRC (Computer Security Resource Center),
NIST.
[483] Samuelson, P. (1990， August). Should program algorithms be patented?. Communications of
the ACM, 33(8)， 23-27.
[484] Samuelson, P. (2004, June). Why reform the U.S. patent system?. Communications of the
ACM, 47(6)， 19-23.
[485] Sandhu, R., deSa, C.， and Uanesan, K. (2003, June 19). One time password entry to access
multiple network sites. USPTO Published Patent Application US2003/01 15452, Filing Date:
19 December 2000.
[486] Sandhu, R., deSa, C, and Ganesan, K. (2005a, April 19). System and method for password
throttling. USPTO Issued Patent US6883095, Filing Date: 19 December 2000, Issue Date: 19
April 2005.
[487] Sandhu, R., deSa, C， and Ganesan, K. (2005b, September 6). High security cryptosystem.
USPTO Issued Patent US6940980, Filing Date: 19 December 2000, Issue Date: 6 September
2005.
309
[488] Sandhu, R.， deSa, C.， and Ganesan, K. (2005c, November 29). System and method for crypto-
key generation and use in crypto system. USPTO Issued Patent US6970562, Filing Date: 19
December 2000, Issue Date: 29 November 2005.
[489] Sandhu, R., deSa, C, and Ganesan, K. (2006a, March 21). Secure communications network
with user control of authenticated personal information provided to network entities. USPTO
Issued Patent US7017041, Filing Date: 19 December 2000, Issue Date: 21 March 2006.
[490] Sandhu, R.， deSa, C.， and Ganesan, K. (2006b, May 30). One time password entry to access
multiple network sites. USPTO Issued Patent US7055032, Filing Date: 21 May 2004, Issue
Date: 30 May 2006.
[491] Sandhu, R.， deSa, C.， and Ganesan, K. (2006c, June 20). System and method for generation
and use of asymmetric crypto-keys each having a public portion and multiple private
portions. USPTO Issued Patent US7065642, Filing Date: 19 December 2000, Issue Date: 20
June 2006.
[492] Sandhu, R.， deSa, C, and Ganesan, K. (2006d, June 27). System and method for
authentication in a crypto -system utilizing symmetric and asymmetric crypto-keys. USPTO
Issued Patent US7069435, Filing Date: 19 December 2000, Issue Date: 27 June 2006.
[493] Sandhu, R.， deSa, C.， and Ganesan, K. (2006e, November 2). Laddered authentication
security using split key asymmetric cryptography. USPTO Published Patent Application
US2006/0248333, Filing Date: 22 June 2006.
[494] Sandhu, R., deSa, C.， and Ganesan, K. (2006f, December 12). Method and system for
authorizing generation of asymmetric crypto-keys. USPTO Issued Patent US7149310, Filing
Date: 19 December 2000, Issue Date: 12 December 2006.
[495] Sandhu, R. S., Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2006a, August 17).
Asymmetric key pair having a kiosk mode. USPTO Published Patent Application
US2006/0 182276, Filing Date: 14 February 2005.
[496] Sandhu, R. S.， Schoppert, B. J., Ganesan, R.， Bellare, M.， and deSa, C. J. (2006b, August 17).
Roaming utilizing an asymmetric key pair. USPTO Published Patent Application
US2006/0 182277, Filing Date: 14 February 2005.
310
[497] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2006c, August 17).
Architecture for asymmetric crypto-key storage. USPTO Published Patent Application
US2006/0 182283, Filing Date: 14 February 2005.
[498] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2006d， August 17).
Technique for asymmetric crypto-key generation. USPTO Published Patent Application
US2006/0 184786, Filing Date: 14 February 2005.
[499] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2006e, August 17).
Authentication protocol using a multi-factor asymmetric key pair. USPTO Published Patent
Application US2006/0 184787, Filing Date: 14 February 2005.
[500] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2006f, August 17).
Multiple factor private portion of an asymmetric key. USPTO Published Patent Application
US2006/0 184788, Filing Date: 14 February 2005.
[501] Sandhu, R. S.， Ganesan, R.， Cottrell, A. P., Renshaw, T. S.， Schoppert, B. J., and Austin, K.
(2007, February 12). Flexible and adjustable authentication in cyberspace. USPTO
Published Patent Application US2007/0199053, Filing Date: 12 February 2007.
[502] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2007a, March 8).
Technique for providing multiple levels of security. USPTO Published Patent Application
US2007/0055878, Filing Date: 14 February 2005.
[503] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2007b, March 22).
Asymmetric crypto-graphy with rolling key security. USPTO Published Patent Application
US2007/0067618, Filing Date: 17 January 2006.
[504] Sandhu, R. S.， Schoppert, B. J.， Ganesan, R.， Bellare, M.， and deSa, C. J. (2007c, November
8). Multifactor split asymmetric crypto-key with persistent key security. USPTO Published
Patent Application US 2007/025 85 85, Filing Date: 5 May 2006.
[505] Sandhu, R. S.， Schoppert, B. J., Ganesan, R., Bellare, M.， and deSa, C. J. (2007d, November
8). Secure login using a multifactor split asymmetric crypto-key with persistent key security.
USPTO Published Patent Application US 2007/025 85 94, Filing Date: 5 May 2006.
[506] Scalet, S. D. (2005, December 1). How to write good passwords, [Online]. CIO.com.
Available: http://www.csoonline.com/article/print/220721 [2008, September 19].
311
[507] Scalise, C. T. (1999). Intellectual property protection reform: Theory, evidence and policy.
Singapore: National University of Singapore, Singapore University Press.
[508] Scarf one, K.， and Dicoi, D. (2007, August 2). Wireless network security for IEEE
802.1 la/b/g and Bluetooth (draft) (NIST Special Publication 800-48 Revision 1 (Draft)).
Gaithersburg, MD， USA: CSRC (Computer Security Resource Center), NIST.
[509] Schiffman, M. D. (2003). Building open source network security tools: Components and
techniques. Indianapolis, Indiana, USA: Wiley Publishing Inc.
[510] Schmeh, K. (2001). Cryptography and public key infrastructure on the Internet. Chichester,
West Sussex, UK: John Wiley & Sons Ltd. (Original work published 2001 in German
language).
[511] Schneider, J. (2004, December 9). Graphical event-based password system. USPTO
Published Patent Application US2004/0250138, Filing Date: 18 April 2003.
[512] Schneier, B. (1996). Applied cryptography: Protocols, algorithms, and code in C (2nd ed.)-
New York City, New York, USA: John Wiley & Sons.
[513] Schneier, B. (2000). Secrets & lies: Digital security in a networked world. New York, NY,
USA: John Wiley & Sons.
[514] Schneier, B. (2006, December 14). MySpace passwords aren't so dumb, [Online]. Wired
News. Available:
http://www.wired.com/politics/security/commentary/securitymatters/2006/12/72300 [2008,
June 11].
[515] Schneier, B. (2007, January 11). Secure passwords keep you safer, [Online]. Wired News.
Available:
http://www.wired.com/politics/security/commentary/securitymatters/2007/01/72458 [2008,
June 11].
[516] Sejarah Melayu [The Malay Annals]. (1954). Singapore: Malaya Publishing House Ltd.
(Original work published 1896 in Jawi language).
[517] Setty, N.， and Gentry, R. L. (2002). Inventorship - A practical guide to determining correct
inventorship, [Online]. Needle & Rosenberg, Atlanta, Georgia. Available:
http://www.needlerosenberg . com/Library/w 1 2^248. doc [2008, July 18].
312
[518] Shannon, C. E. (1948, July & October). A mathematical theory of communication. Bell
System Technical Journal, 27， 379-4ZJ & 623-656.
[519] Shannon, C. E. (1949， October). Communication theory of secrecy systems. Bell System
Technical Journal, 28, 65o- /
[520] Shannon, C. E. (1951, January). Prediction and entropy of printed english. Bell System
Technical Journal, 30(1), 50-64.
[521] Shellabear, W. G. (1975). Sejarah Me lay u [The Malay Annals]. Kuala Lumpur, Malaysia:
Penerbit Fajar Bakti Sdn. Bhd. (In Malaysian language).
[522] Shelton, R. H. (2007, March 29). System and method of licensing intellectual property assets.
USPTO Published Patent Application US 2007/0073625, Filing Date: 27 September 2005.
[523] Siegel, L. J. (2005, March 1). Criminology (9th ed.). Florence, KY， USA: Cengage Learning,
Inc., Wadsworth Publishing (URL: http://www.wadsworth.com) (URL:
http://www.cengage.com).
[524] Silverman, J. H. (1986). The arithmetic of elliptic curve. New York, NY, USA: Springer-
Verlag New York, Inc.
[525] Simmons, G. J. (1984， April 9-11). The subliminal channel and digital signatures.
Proceedings of the Advances in Crypto logy (EUROCRYPT '84) (aka Proceedings of the
Workshop on the Theory and Application of Cryptographic Techniques 1984), LNCS 209,
Paris, France, 364-378.
[526] Simmons, G. J. (1998， May). The history of subliminal channels. IEEE Journal on Selected
Areas in Communications, 7(5(4), 452-462.
[527] 新加坡 女郎 患 湿 疫无指 纹纹机 场通关 检查花 45 分 鐘 [Singaporean Female having eczema
has no nngerprint spending 45 minutes to get through the airport immigration office]. (2008,
August 22-23). 星 洲 曰 报 [Sin Chew Daily], [Online], p. 17. Available:
http://search.sinchew-i.com/node/192066 [2008, September 1].
[528] Singh, S. (1999). The code book: The secret history of codes & code-breaking. London, UK:
Fourth Estate Limited, 1-32, 79-80.
313
[529] Singh al, A., Winograd, T.， and Scarfone, K. (2007, August). Guide to security web services
(NIST Special Publication 800-95). Gaithersburg, MD， USA: CSRC (Computer Security
Resource Center), NIST.
[530] Siu, H. S., Chin, Y. H.， and Yang, W. P. (1998a, April). Byzantine agreement in the presence
of mixed faults on processors and links. IEEE Transaction on Parallel and Distributed
Systems, 9(4)， 335-345.
[531] Siu, H. S., Chin, Y. H.， and Yang, W. P. (1998b, July). Reaching strong consensus in the
presence of mixed failure types. Information Sciences, 108{\-4), 157-180.
[532] Smith, A. J. (1990， April). The task of the referee. IEEE Computer, 23(4), 65-71.
[533] Spafford, E. H. (1992a, May). Opus: Preventing weak password choices. Computers &
Security, 77(3), 273-278.
[534] Spafford, E. H. (1992b, September 14-17). Observations on reusable password choices.
Proceedings ofUSENIX3rd Symposium on UNIX Security, Baltimore, MD， USA, 299-312.
[535] Spafford, E. H. (1993). What is a Ph.D. dissertation? , [Online]. Available:
http://spaf.cerias.purdue.edu/〜spa^Archive/spaf.html [2008, August 2].
[536] Spours, P. (2006, April). How to exploit patents for profit, [Online]. IP Review, 14. Available:
http://wwwxpaglobalxom/ip-review-online/1221/how_to_exploit_patents_for_profit [2008,
September 3].
[537] Stallings, W. (1995). Protect your privacy: A guide for PGP users. Englewood Cliffs, NJ,
USA: Prentice Hall.
[538] Stallings, W. (2000). Network security essentials: Applications and standards. Upper Saddle
River, New Jersey, USA: Prentice Hall.
[539] Stallings, W. (2005). Wireless communications & networks (2nd ed.)- Upper Saddle River,
NJ, USA: Pearson Prentice Hall.
[540] Stallings, W. (2006a). Computer organization and architecture (7th ed.). Upper Saddle River,
NJ, USA: Pearson Prentice Hall.
314
[541] S tailings, W. (2006b). Cryptography and network security: Principles and practices (4th ed.).
Upper Saddle River, NJ, USA: Pearson Prentice Hall.
[542] S tailings, W. (2007). Data and computer communications (8th ed.)- Upper Saddle River, NJ,
USA: Pearson Prentice Hall.
[543] Standing, L. (1973， May). Learning 10,000 pictures. Quarterly Journal of Experimental
Psychology, 25(2), 207-222.
[544] Standing, L.， Conezio, J.， and Haber, R. (1970). Perception and memory for pictures: Single-
trial learning of 2500 visual stimuli. Psychonomic Science, 79(2), 73-74.
[545] S tin son, D. R. (2002). Cryptography: Theory and practice. Boca Raton, FL， USA: Chapman
& Hall/CRC Press.
[546] Stubblefield, N. B. (1908， May 12). Wireless telephone. USPTO Issued Patent US887357,
Filing Date: 5 April 1907， Issue Date: 12 May 1908.
[547] Suo, X.， Zhu, Y.， and Owen, G. S. (2005, December 5-9). Graphical passwords: A survey.
Proceedings of the 21st Annual Computer Security Applications Conference 2005 (ACSAC
2005), Tucson, Arizona, USA, 463-472.
[548] Swanson, M. D.， Kobayashi, M.， and Tewfik, A. H. (1998， June). Multimedia data-
embedding and watermarking techniques. Proceedings of the IEEE, 86(6), 1064-1087.
[549] Sweet, W. (2007, November). Keeping score in the IP game. IEEE Spectrum, 44(11)， 39-41.
[550] Tai, C. L. (2001, December 2^;. Dual in-line type finger-actuated switch. USPTO Issued
Patent US6333479, Filing Date: 4 January 2001， Issue Date: 25 December 2001.
[551] Tan, C. P. (1981, May). On the entropy of the Malay language. IEEE Transactions on
Information Theory, 27(3)， 383-384.
[552] Tan, C.-P., and Yap, S.-T. (2001, November 1). Small-sample studies of the entropy of
Chinese (Tech. Rep. No. 6/2000). Kuala Lumpur, Malaysia: University of Malaya, Institute
of Mathematical Sciences.
315
[553] Taplin, R. (2004, October). Protect and survive: Managing intellectual property in the Far
East - The case of Japan, [Online]. Thomson Scientific. Available:
http://scientific.thomson.com/free/ipmatters/bti/8249985/ [2007, May 16].
[554] Tareq Rajab Museum (TSR). (1998-2002). Manuscripts in the TSR Museum, [Online]. Tareq
Rajab Museum, Hawelli, Kuwait. Available: http://www.trmkt.com/manu.html [2008, July
16].
[555] Tavakoli, N. (1991， June 24-28). Information content oi images. Proceedings of the 1991
IEEE International Symposium on Information Theory, Budapest, Hungary, 264.
[556] Taylor, N. (2002). Laser: The inventor, the Nobel laureate, and the thirty-year patent war.
New York City, NY, USA: Simon & Schuster.
[557] Thatcher, M. E.， and Pingry, D. E. (2007, October). [Software patents] The good, the bad,
and the messy. Communications of the ACM, 50(10)， 47-52.
[558] 圣经 (新约 全书) (中 英对 照) (和合 本) [The Holy Bible (New Testament) (Chinese and
English bilingualed.) (New King James Version)]. (199/ ノ . Nashville, Tennessee, USA:
Thomas Nelson, & Taichung City [台 中 市]， Taiwan (ROC) [中華 民國臺 灣]: The Gideons
International (Taiwan (ROC)) [國際 基甸會 中華 民國總 會]. (in Chinese and English
languages).
[559] The official Scrabble players dictionary (3rd ed.)- (2002). Heatherton, Victoria, Australia:
Hinkler Books.
[560] The proof of Parkinson, [Online]. (1969， July 18). Time Magazine. Available:
http://www.timexom/time/magazine/article/0,9171,901078,00.html [2008, July 18].
[561] The Star Online. (2007, June 8). Exchange rate, [Online]. Available:
http://biz.thestar.com.my/business/exchange.asp [2007, June 8].
[562] The Unicode Consortium, Allen, J. D.， et al. (Eds.). (2006, November 19). The Unicode
Standard 5.0. Boston, Massachusetts, USA: Addison-Wesley Professional.
[563] The United Nations Convention against transnational organized crime and its protocols,
[Online]. (2004). UNODC (United Nations Office on Drugs and Crime). Available:
http://www.unodc.org/unodc/en/treaties/CTOC/index.html;
316
http://www. 皿 odc.org/documents/treaties/raTOC/Publications/T^
book-e.pdf [2008, July 21].
[564] Thomas, P., and Breitzman, A. (2006a, January 12). The influence of IEEE on key patents,
[Online]. IEEE. Available:
http://www.ieee.org/portal/cms_docs_iportals/discover/sub_pages/IEEE_Key_Patents_2006.
pdf [2008, September 3].
[565] Thomas, P., and Breitzman, A. (2006b, August 1). A method for identiiymg hot patents and
linking them to government-funded scientific research. Research Evaluation, 75(2), 145-152.
[566] Thorpe, J.， and van Oorschot, P. C. (2007, August 6-10). Human- seeded attacks and
exploiting hot-spots in graphical passwords. Proceedings of the 16th US EN IX Security
Symposium 2007 (Security 2007), Boston, MA, USA, 103-118.
[567] Tomasi, W. (1998). Electronic communications systems: Fundamentals through advanced
(3rd ed.). Upper Saddle River, NJ: Prentice-Hall.
[568] Tracy, M.， Jan sen, W.， Scarf one, K., and Butterfield, J. (2007, February). Guidelines on
electronic mail security (NIST Special Publication 800-45 Version 2). Gaithersburg, MD，
USA: CSRC (Computer Security Resource Center), NIST.
[569] Tracy, M.， Jansen, W.， Scarfone, K.， and Winograd, T. (2007, September). Guidelines on
securing public web servers (NIST Special Publication 800-44 Version 2). Gaithersburg, MD，
USA: CSRC (Computer Security Resource Center), NIST.
[570] Tsou, B. K.， Lai, T. B. Y.， and Chow, K.-P. (2004, March 22-24). Comparing entropies
within the Chinese language. Proceedings of 1st International Joint Conference on Natural
Language Processing (IJCNLP 2004), LNCS 3248, Hainan Island, China, 466-475.
[571] Turk, S. A. (2005, Spring). The most important 21st century intellectual property issue.
Chicago-Kent Journal of Intellectual Property, 4(2), 153-155.
[572] Twelve steps to developing an effective first draft of your manuscript, [Online]. (No date).
San Francisco Edit. Available: http://www.cis.nctu.edu.tw/~tzeng/First%20Draftpdf [2008,
August 2].
317
[573] UKCS. (No date a). Commonly quoted 'alternatives' to registration, [Online]. The UK
Copyright Service (UKCS). Available:
http://www.copyrightservice.co.uk/services/alternatives [2007, May 20].
[574] UKCS. (No date b). The registration service, [Online]. The UK Copyright Service (UKCS).
Available: http://www.copyrightservice.co.uk/register/ [2007, June 7].
[575] United Publishing House [联营 出版 有限 公司]. (2001). 最新 汉语 大词典 [修订 版] [The
newest Chinese language big phrase dictionary (revised ed.)]- Seri Kembangan, Selangor,
Malaysia: United Publishing House, (in Chinese language).
[576] United Publishing House [联营 出版 有限 公司]. (2002). 新汉语 字典 [New Chinese
language word/character diction arvj. Sen Kemoangan, Selangor, Malaysia: United
Publishing House, (in Chinese language).
[577] UNODC and organized crime, [Online]. (No date). UNODC (United Nations Office on
Drugs and Crime). Available: http://www.unodc.org/imodc/en/organized-crime/index.html
[2008, July 23].
[578] U.S. Department of Defense. (1985). Password management guideline (Report No. CSC-
STD-002-85) Fort George G. Meade, Maryland, USA: DoD Computer Center.
[579] USCO. (No date). Current fees (July 2007 - j, [Online]. US Copyright Office (USCO).
Available: http://www.copyright.gov/docs/fees.html [2008, May 20].
[580] Recommendations for consideration by the incoming administration regarding the US Patent
and Trademark Office, [Online]. (No date). US Chamber of Commerce (USCOC). Available:
http://dotank.nyls.edu/communitypatent/docs/USCOCRecommendations.pdf [2008,
November 10].
[581] Vacca, J. R. (2007, March 16). Biometric technologies and verification systems. Oxford,
Oxfordshire, UK: Butterworth-Heinemann, p. 280.
[582] Vines, P., and Zobel, J. (1998, October). Compression techniques for Chinese text. Software:
Practice and Experience, 25(12), 1299-1314.
[583] Wag staff Jr., S. S. (2003). Cryptanalysis of number theoretic ciphers. Boca Raton, FL， USA:
Chapman & Hall/CRC Press, 115-117.
318
[584] Wailgum, T. (2008, September 8). Password brain teaser: Too many passwords or not
enough brain power?, [Online]. CIO.com. Available:
http://www.cio.com/article/print/448241 [2008, September 10].
[585] Walker, P. B. (2007, November 8). Intangible property transaction and leaseback business
method. USPTO Published Patent Application US 2007/0260549, Filing Date: 4 May 2006.
[586] Wang, L. [汪 力] (Ed.). (2007, August). 有组织 犯罪 专题 研究 [Research on the topic of
organized crime]. Beijing [北京]， China [中国]: People's Publishing House [人民 出版社],
(in Chinese language).
[587] Wang, S. C， and Kao, S. H. (2001, January 31 - February 2). A new approach for Byzantine
agreement. Proceedings of the 15th International Conference on Information Networking
2001 (ICOIN2001), Beppu City, Oita, Japan, 518-524.
[588] Wang, S. C.， and Yan, K. Q. (2000, July 4-7). Reaching fault diagnosis agreement on dual
link failure mode. Proceedings of the 7th International Conference on Parallel and
Distributed Systems 2000 (ICPADS 2000), Iwate, Japan, 291-298.
[589] Wang, X.， Yin, Y. L.， and Yu， H. (2005, August 14-18). Finding collisions in the full SHA-1.
Proceedings of the 25th Annual International Cryptology Conference on Advances in
Cryptology (Crypto 2005), LNCS 3621， Santa Barbara, CA， USA, 17-36.
[590] Wang, X. L. [王 学理]， and Pei, D， Y. [裴定 一]. (2006). 摘 圆与超 摘 圆曲线 公钢密 码的理
论ョ 头现 [Theory and implementation of elliptic curve and hyperelliptic curve
cryptography]. Beijing [北京]， China [中国]: Science Press [科学 出版社], (in Chinese
language).
[591] Wangerin, W. Jr. [沃 爾特。 溫傑 林] (1999a). The book of God: New Testament [小 說聖
經: 新約 篇]. (Song，B.-Y. [宋碧 雲]， Trans,). Taipei City [臺北 市]， Taiwan (ROC) [中華
民國臺 灣]: Taiwan Wisdom Publishing Co., Ltd. [臺 灣先智 出版 事業 股份 有限 公司], (in
Chinese language;.
[592] Wangerin, W. Jr. [沃 爾特。 溫傑 林] (1999b). The book of God: Old Testament [小說 聖經:
舊約 篇]. (Gao, Z.-R. [高 志 仁]， Trans.). Taipei City [臺北 市]， Taiwan (ROC) [中華 民國臺
灣]: Taiwan Wisdom Publishing Co., Ltd. [臺 灣先智 出版 事業 股份 有限 公司], (in Chinese
language).
319
[593] Weinshall, D.， and Kirkpatrick, S. (2004, April 24-29). Passwords you'll never forget, but
can't recall. Proceedings of the Conference on Human Factors in Computing Systems 2004
(CHI 2004), Vienna, Austria, 1399-1402.
[594] Weiss, R. P., and South, N. (Eds.)- (1998). Comparing prison systems: Toward a
comparative and international penology. Amsterdam, The Netherlands: OP A (Overseas
Publishers Association) under the license from GIB (Gordon and Breach Publishers).
[595] Wiener, M. J. (1990， May). Cryptanalysis of short RSA secret exponents. IEEE Transactions
on Information Theory, 36(3), 553-558.
[596] Wikipedia Contributors. (2006, March 14). Statutory damages, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Statutory_damages&oldid=43705331 [2007, May
23].
[597] Wikipedia Contributors. (2007a, January 13). Four corner method, [Online]. Wikipedia The
Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Four_corner_method&oldid=100425004 [2007,
June 2].
[598] Wikipedia Contributors. (2007b, February 10). Password policy, [Online]. Wikipedia The
Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Password_policy&oldid=106996972 [2007, June
2].
[599] Wikipedia Contributors. (2007c, February 26). Chinese input methods for computers,
[Online]. Wikipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Chinese_inpu ᄂ methods_for_computers&oldid=ll
1110951 [2007, June 2].
[600] WiKipedia Contributors. (2007d, March 17). Chinese character encoding, [Online].
WiKipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Chinese_character_encoding&oldid= 115862258
[2007, June 2].
[601] WiKipedia Contributors. (2007e, March 31a). Statutory damages for copyright infringement,
[Online]. Wikipedia The Free Encyclopedia. Available:
320
http://en.wikipedia.org/w/index.php?title=Statutory_damages_fo ᄂ
oldid= 11 9234996 [2007, May 23].
[602] Wikipedia Contributors. (2007f, March 31b). Typeface, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Typeface&oldid=119171254 [2007, June 2].
[603] Wikipedia Contributors. (2007g, April 5). ASCII art, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=ASCII_art&oldid=120415207 [2007, June 2].
[604] Wikipedia Contributors. (2007h, April 5). Chinese character, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Chinese_character&oldid= 1205 19979 [2007, June
2].
[605] WiKipedia Contributors. (2007i, June 6). Berne Convention for the Protection of Literary and
Artistic Works, [Online]. Wikipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Beme_Convention_for_the_Protection_of_Literar
y_and_Artistic_Works&oldid=136381222 [2007, June 7].
[606] Wikipedia Contributors. (2007j, September 13). First to file and first to invent, [Online].
Wikipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Firs ᄂ to_file_and_firs ᄂ to_inve^
86 [2007, October 22].
[607] Wikipedia Contributors. (2007k, September 20). TIA/E1A-568-B, [Online]. Wikipedia The
Free Encyclopedia. Available: http://en.wikipedia.Org/w/index.php ?title=TIA/EIA-568-
B&oldid=159241053 [2008, January 24].
[608] Wikipedia Contributors. (20071, September 26). Registered jack, [Online]. Wikipedia The
Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Registered_jack&oldid=160450633 [2008, January
24].
[609] Wikipedia Contributors. (2007m, September 27). Modular connector, [Online]. Wikipedia
The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Modular_connector&oldid= 160734170 [2008,
January 24].
321
[610] Wikipedia Contributors. (2007n, October 4). Gordon Gould, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Gordon_Gould&oldid=162298481 [2007, October
10].
[611] Wikipedia Contributors. (2007o, October 6). Power over Ethernet, [Online]. Wikipedia The
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Power_over_Ethernet&oldid=162586392 [2008,
January 24].
[612] Wikipedia Contributors. (2007p, October 7). Category 5 cable, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Category_5_cable&oldid=162908384 [2008,
January 24].
[613] Wikipedia Contributors. (2007q, October 7). Hacker (computer security), [Online].
Wikipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Hacker_%28computer_security%29&oldid= 16295
3161 [2008, January 24].
[614] Wikipedia Contributors. (2007r, December 2). Penology, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Penology&oldid=175222217 [2008, July 23].
[615] Wikipedia Contributors. (2008a, March 6). Monolingualism, [Online]. Wikipedia The Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Monolingualism&oldid=196315794 [2008, July
23].
[616] Wikipedia Contributors. (2008b, March 8). Patent Reform Act of 2005, [Online]. Wikipedia
The Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Patent_Reform_Act_of_2005&oldid=197359847
[2008, May 20].
[617] Wikipedia Contributors. (2008c, March 10). Linguistic demography, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Linguistic_demography&oldid= 197 188475 [2008,
July 23].
322
[618] Wikipedia Contributors. (2008d, April 26). Ocean Tomo LLC, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Ocean_Tomo_LLC&oldM [2008,
August 21].
[619] Wikipedia Contributors. (2008e, May 5). Patent Reform Act of 2007, [Online]. Wikipedia
The Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Patent_Reform_Act_of_2007&oldid=210421848
[2008, May 20].
[620] Wikipedia Contributors. (2008f, May 7). Redundancy (information theory), [Online].
Wikipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Redundancy_%28informatior ᄂ theory%29&oldid=
210871144 [2008, May 20].
[621] Wikipedia Contributors. (2008g, May 15). CamelCase, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=CamelCase&oldid=21268m [2008, May 20].
[622] Wikipedia Contributors. (2008h, May 30). Theoretical linguistics, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=TheoreticalJinguistics&oldid=216055905 [2008,
July 23].
[623] Wikipedia Contributors. (2008i, June 21). Coefficient of inefficiency, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Coefficien ᄂ of_Inefficiency&oldid=2209695 34
[2008, July 23].
[624] Wikipedia Contributors. (2008j, June 21). Language policy, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Language_policy&oldid=220714548 [2008, July
23].
[625] Wikipedia Contributors. (2008k, June 22). Petname, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Petname&oldid=220857363 [2008, July 16].
323
[626] Wikipedia Contributors. (20081, July 3). Short-term memory, [Online]. AVikipedia the Free
Encyclopedia. Available: http://en. wikipedia. org/w/index.php?title=Short-
term_memory&oldid=223406253 [2008, July 23].
[627] Wikipedia Contributors. (2008m, July 7). List of writing systems, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=List_of_writing_systems&oldid=224239367 [2008,
September 1].
[628] Wikipedia Contributors. (200 8n, July 8). Timeline of computer security hacker history,
[Online]. Wikipedia The Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Timeline_of_computer_security_hacker_history&
oldid=224341103 [2008, July 16].
[629] Wikipedia Contributors. (2008o, July 9). Cryptographic hash function, [Online]. AVikipedia
the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Cryptographic_hash_function&oldid=224604390
[2008, July 17].
[630] Wikipedia Contributors. (2008p, July 9). List of multilingual countries and regions, [Online].
Wikipedia the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Lis ᄂ of_multilingua ᄂ countries_and_regions&oldi
d=224607350 [2008， July 23].
[631] Wikipedia Contributors. (2008q, July 11). Cronyism, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Cronyism&oldid=225^ [2008, July 23].
[632] Wikipedia Contributors. (2008r, July 11). MD5, [Online]. Wikipedia the Free Encyclopedia.
Available: http://en.wikipedia.org/w/index.php?tifle=MD5&oldid= 225 () 4 l 24 4 [2008, July
16].
[633] Wikipedia Contributors. (2008s, July 12). Criminology, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Criminology&oldid=225212774 [2008, July 23].
[634] Wikipedia Contributors. (2008t, July 13). Dissertation, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Dissertation&oldid=225330900 [2008, July 14].
324
[635] Wikipedia Contributors. (2008u, July 15). Applied linguistics, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Applied_linguistics&oldid=225753837 [2008, July
23].
[636] Wikipedia Contributors. (2008v, July 15). Moore's Law, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia,org/w/index.php?title=Moore%27s_law&oldid=^ [2008, July
16].
[637] Wikipedia Contributors. (2008w, July 15). SHA hash functions, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=SHA_hash_functions&oldid=225763377 [2008,
July 16].
[638] Wikipedia Contributors. (2008x, July 17). Parkinson's Law, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Parkinson%27s_law&oldid=226138504 [2008,
July 23].
[639] Wikipedia Contributors. (2008y, July 19). Corpus linguistics, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Corpus_linguistics&oldid [2008, July
23].
[640] Wikipedia Contributors. (2008z, July 19). Text corpus, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Text_corpus&oldid=226566944 [2008, July 23].
[641] Wikipedia Contributors. (2008aa, July 21). First language, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=First_language&oldid=227072714 [2008, July 23].
[642] Wikipedia Contributors. (2008ab, July 21). Second language, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Second_language&oldid=226939389 [2008, July
23].
325
[643] Wikipedia Contributors. (2008ac, July 22). Ethnologue list of most spoken languages,
[Online]. Wikipedia the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Ethnologue_lis ᄂ of_mos ᄂ spoken_languages&oldi
d=227286397 [2008, July 23].
[644] Wikipedia Contributors. (2008ad, July 22). List of languages by number of native speakers,
[Online]. Wikipedia the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=List_of_languages_by_number_of_native_speaker
s&oldid=227300820 [2008, July 23].
[645] Wikipedia Contributors. (2008ae, July 22). Multilingualism, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Multilingualism&oldid=227240877 [2008, July
23].
[646] Wikipedia Contributors. (2008af, July 22). Organized crime, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Organized_crime&oldid=227261189 [2008, July
23].
[647] Wikipedia Contributors. (2008ag, July 22). SMS language, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=SMS_language&oldid=22^ [2008, July 23].
[648] Wikipedia Contributors. (2008ah, July 22). Text messaging, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Text_messaging&oldid=227309098 [2008, July
23].
[649] Wikipedia Contributors. (2008ai, July 23). Language, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Language&oldid=227357180 [2008, July 23].
[650] Wikipedia Contributors. (2008aj, July 23). Sans-serif, [Online]. Wikipedia the Free
Encyclopedia. Available: http://en.wikipedia.org/w/index.php?title=Sans-
serif&oldid=227325014 [2008, July 23].
326
[651] Wikipedia Contributors. (2008ak, July 23). Serif, [Online]. Wikipedia the Free Encyclopedia.
Available: http://en.wikipedia.org/w/index.php?title=Serif&oldid=22734154^ [2008， July
23].
[652] Wikipedia Contributors. (2008al, August 9). Digital identity, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Digital_identity&oldid=230770248 [2008， August
25].
[653] Wikipedia Contributors. (2008am, August 11). OpenlD, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=OpenID&oldid=231207285 [2008, August 16].
[654] Wikipedia Contributors. (2008an, August 12). Alcatel-Lucent v. Microsoft, [Online].
Wikipedia the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Alcatel-Lucen ᄂ v._Microsoft&oldid=23 1503895
[2008, September 17].
[655] Wikipedia Contributors. (2008ao, August 15). Inventor (patent), [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Inventor_(patent)&oldid=232136205 [2008,
August 16].
[656] Wikipedia Contributors. (2008ap, August 19). Information Card, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Information_Card&oldid=233004113 [2008,
August 23].
[657] Wikipedia Contributors. (2008aq, August 27). Writing system, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Writing_system&oldid=234534887 [2008,
September 1],
[658] Wikipedia Contributors. (2008ar, September 8). List of languages by total number of
speakers, [Online]. Wikipedia the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=List_of_languages3y_total_number_of_speakers
&oldid=237048814 [2008， October 4].
327
[659] Wikipedia Contributors. (2008as, September 18). Long-term memory, [Online]. Wikipedia
the Free Encyclopedia. Available: http://en. wikipedia. org/w/index.php?title=Long-
tera 니 nemory&oldid=239249968 [2008, October 4].
[660] Wikipedia Contributors. (2008at, September 23). CJK Unified Ideographs, [Online].
Wikipedia the Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=CJK_Unified_Ideographs&oldid=240390843
[2008, October 16].
[661] Wikipedia Contributors. (2008au, September 27). Languages by speakers, [Online].
Wikipedia the Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Languages_by_speakers&oldid=241385108 [2008,
October 4].
[662] Wikipedia Contributors. (2008av, October 3). Multilingual education, [Online]. AVikipedia
the Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Multilingual_Education&oldid=242699632 [2008,
October 4].
[663] Wikipedia Contributors. (2008aw, October 4). Working memory, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Working_memory&oldid=242892307 [2008,
October 4].
[664] Wikipedia Contributors. (2008ax, October 9). Spy satellite, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Spy_satellite&oldid=244048302 [2008, October
11]-
[665] Wikipedia Contributors. (2008ay, October 10). GeoEye, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=GeoEye&oldid= 24444 3 5 2^ [2008, October 11].
[666] Wikipedia Contributors. (2008az, October 10). MIFARE, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia,org/w/index.php?title=MIFARE&oldid=24438692^ [2008, October 16].
[667] Wikipedia Contributors. (2008ba, October 12). Unity of invention, [Online]. Wikipedia the
Free Encyclopedia. Available:
328
http://en.wikipedia.org/w/index.php?title=Unity_of_invention&oldid=244841923 [2008.
October 15].
[668] Wikipedia Contributors. (2008bb, October 14). Age of Discovery, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Age_of_Discovery&oldid=245322423 [2008,
October 16].
[669] Wikipedia Contributors. (2008bc, October 15). British Empire, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=British_Empire&oldid=245420209 [2008, October
16].
[670] Wikipedia Contributors. (2008bd, October 23). Botriet, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Botnet&oldid=247128093 [2008, October 23].
[671] Wikipedia Contributors. (2008be, October 23). Disk cloning, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Disk_cloning&oldid=247128872 [2008, October
23].
[672] Wikipedia Contributors. (2008bf, October 23). Disk image, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Disk_image&oldid=247128569 [2008, October
23].
[673] Wikipedia Contributors. (2008bg, October 23). H-index, [Online]. Wikipedia the Free
Encyclopedia. Available: http://en. wikipedia. org/w/index.php?title=H-
index&oldid=247127081 [2008, October 23].
[674] Wikipedia Contributors. (2008bh, October 23). Identity theft, [Online]. Wikipedia the Free
Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Identity_theft&oldid=246961579 [2008, October
23].
[675] Wikipedia Contributors. (2008bi, October 23). Internet bot, [Online]. Wikipedia the Free
Encyclopedia. Available:
329
http://en.wikipedia.org/w/index.php?title=Internet_bot&oldid=247128048 [2008, October
23].
[676] Wikipedia Contributors. (2008bj, October 23). Zombie computer, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Zombie_computer&oldid=247 127637 [2008,
October 23].
[677] Wikipedia Contributors. (2008bk, October 25). Extrasensory perception, [Online]. AVikipedia
the Free Encyclopedia. Available:
http://en.wikipedia.Org/w/index.php ?title=Extrasensory_perception&oldid=247523584 [2008,
October 25].
[678] Wikipedia Contributors. (2008bl, October 25). Psi (parapsychology), [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Psi_(parapsychology)&oldid=2475 24054 [2008,
October 25].
[679] Wikipedia Contributors. (2008bm, October 25). Parapsychology, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title=Parapsychology&oldid=247525788 [2008,
October 25].
[680] Wikipedia Contributors. (2008bn, October 25). Parapsychological Association, [Online].
Wikipedia the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Parapsychological_Association&oldid=2475 24424
[2008, October 25].
[681] Wikipedia Contributors. (2008bo, October 25). Joseph Banks Rhine, [Online]. Wikipedia the
Free Encyclopedia. Available:
http://en.wikipedia.org/w/index.php?title= Joseph Janks_Rhine&oldid=247524655 [2008,
October 25].
[682] Wikipedia Contributors. (2008bp, October 25). List of psychic abilities, [Online]. Wikipedia
the Free Encyclopedia. Available:
http://en. wikipedia. org/w/index.php?title=Lis ᄂ of_psychic_abilities&oldid=247524868
[2008, October 25].
330
[683] Williams, A. (2006, February 16). System and method for patent evaluation using artificial
intelligence. USPTO Published Patent Application US2006/0036632, Filing Date: 11 August
2004.
[684] Williams, L. C. (2002). A discussion of the importance of key length in symmetric and
asymmetric cryptography, [Online]. Bethesda, Maryland, USA: SANS Institute. Available:
http://www.giac.org/practical/gsec/Lorraine_Williams_GSEC.pdf [2008, May 17].
[685] Wilson, A. L. (2008, June 11). Microsoft's CardSpace attacked by researchers, [Online].
CIO.com. Available: http://www.cio.com/article/print/391813 [2008, September 10].
[686] Wilson, C.， Grother, P., and Chandramouli, R. (2007, January). Biometric data specification
for personal identity verification (NIST Special Publication 800-76-1). Gaithersburg, MD，
USA: CSRC (Computer Security Resource Center), NIST.
[687] Windows Live ID (aka Microsoft Passport Network), [Online]. (No date). Available:
http://www.passport.net [2008, July 16].
[688] Wingate, P., and Sinden, D. (Eds.). (2002, September). Totltxt: The big Book of little text
messages. New York, NY, USA: Book Sales, Inc. (URL: http://www.booksalesusa.com).
[689] WIPO. (No date). Pre-established damages I statutory damages. In Which kind of damages
are available in IP disputes?, [Online]. WIPO. Available:
http://www.wipo.in t/enforcement/en/faq/judiciary/faq08.html#pre [2007, May 23].
[690] WIPO. (1979， September 28). Berne Convention for the Protection of Literary and Artistic
Works, [Online]. WIPO. Available:
http://www.wipo.int/treaties/en/ip/berne/trtdocs_wo001.html [2007, May 20].
[691] Witty, R. J. (2001, January 30). Best practices for managing PINs and passwords (Tech. Rep.
No. Gartner QA- 12-8664). Stamford, CT， USA: Gartner, Inc.
[692] Wolfgang, R. B.， Podilchuk, C. L， and Delp, E. J. (1999， July). Perceptual watermarks for
digital images and video. Proceedings of the IEEE, 87(7), 1108-1126.
[693] Wrixon, F. B. (1998). Codes ciphers & other cryptic & clandestine communication: Making
& breaking secret messages from hieroglyphs to the Internet. New York City, NY, USA:
Konemann, 298-303, 671-672.
331
[694] Wu， T. J. (2003, March 25). System and method for securely logging onto a remotely located
computer. USPTO Issued Patent US6539479, Filing Date: 14 July 1998， Issue Date: 25
March 2003.
[695] Wu, X.， Cheng, S.， and Xiong, Z. (2001, March). On packetization of embedded multimedia
bitstreams. IEEE Transactions on Multimedia, 3(1), 132-140.
[696] Xu， F. C. [徐梵 澄]. (1984, January). S"hMXH [Fifty Upanishads]. Beijing, China: China
Social Sciences Press [中国 社会 禾斗 누 出版社], (in Chinese language).
[697] Xu， S. [许 慎]. (2001). 说文 解字 [Talking on language and explaining the words]. Hong
Kong SAR, China: Chung Hwa Book [中 华书 局], (in Chinese language).
[698] Yan, J.， Blackwell, A.， Anderson, R.， and Grant, A. (2004, September-October). Password
memorability and security: Empirical results. IEEE Security and Privacy Magazine, 2(5), 25-
31.
[699] Yan, K. Q.， and Chin, Y. H. (1988， August). An optimal solution for consensus problem in
an unreliable communication system. Proceedings of the International Conference on
Parallel Processing (ICPP 1988), The Pennsylvania State University, University Park, PA,
USA, Vol. 1: Architecture, 388-391.
[700] Yan, K. Q.， Chin, C. H.， and Wang, S. C. (1992, June). Optimal agreement protocol in
malicious faulty processors and faulty links. IEEE Transactions on Knowledge and Data
Engineering, 4(3), 266-280.
[701] Yan, K. Q.， and Wang, S. C. (2005a, February 25). Reaching fault diagnosis agreement on an
unreliable general network. Information Sciences, 1 70(2-4), 397-407.
[702] Yan, K. Q.， and Wang, S. C. (2005b, July). Grouping Byzantine agreement. Computer
Standards & Interfaces, 28(1), 75-92.
[703] Yan, K. Q.， Wang, S. C， and Chin, Y. H. (1999， March 12). Consensus under unreliable
transmission. Information Processing Letters, (59(5), 243-248.
[704] Yannakoudakis, E. J., and Angelidakis, G. (1988， November). An insight into the entropy
and redundancy of the English dictionary. IEEE Transactions on Pattern Analysis and
Machine Intelligence, 10(6), 960-970.
332
[705] Yee, K. P., and Sitaker, K. (2006, July 12-14). Passpet: Convenient password management
and phishing protection. Proceedings of Symposium on Usable, Privacy and Security
(SOUPS2006), Pittsburgh, PA, USA, 32-43.
[706] Zhang, H.， Xu， B.， and Huang, T. (2000, October 14-16). Statistical analysis of Chinese
language and language modeling based on huge text corpora. Proceedings of the 3rd
International Conference on Advances in Multimodal Interfaces (I CM I 2000), LNCS 1948，
Beijing, China, 279-286.
[707] Zheng, Y. (1997， August 17-21). Digital signcryption or how to achieve cost(signature &
encryption) « cost(signature) + cost(encryption) . Proceedings of the 1 7th Annual
International Crypto logy Conference on Advances in Crypto logy (CRYPTO ,9T), LNCS
1294, Santa Barbara, CA， USA, 165-179.
[708] Zhu， Y. F. [祝跃 飞]， and Zhang, Y. J. [张亚 娟]. (2006, October). 補 圆曲线 公钢密 码导引
[Guide to elliptic curve cryptography]. Beijing [北京]， China [中国]: Science Press [주4 쫓 1 出
社], (in Chinese language).
333
ACRONYMS
2TDEA
2-Key Triple Data Encryption Algorithm
3TDEA
3-Key Triple Data Encryption Algorithm
2TDES
2-Key Triple Data Encryption Standard
3TDES
3-Key Triple Data Encryption Standard
ACM
Association for Computing Machinery
AES
Advanced Encryption Standard
AIPO/OAPI
African Intellectual Property Organization
(Organisation Africaine de la Propriete Intellectuelle)
ANN
Artificial Neural Network
ANN Based BAP Artificial Neural Network Based Byzantine Agreement Protocol
APA American Psychological Association
APWG Anti-Phishing Working Group
ARIPO African Regional Industrial Property Organization
AUTM Association of University Technology Managers
BAP Byzantine Agreement Protocol
BAP-ANN Byzantine Agreement Protocol with Artificial Neural Network
BGP Byzantine Generals Problem
BTIRDM Budapest Treaty on the International Recognition of the Deposit of Microorganisms
for the Purposes of Patent Procedure
CAPTCHA Completely Automated Public Turing test to tell Computers and Humans Apart
CII Computer-Implemented Invention
CIS Cryptography & Information Security
CLJ Crime, Law, and Justice
CLPP Chinese Language Passphrase
CLPW Chinese Language Password
CM Communication Management
CO Central Office
CSPRNG Cryptographically Secure Pseudo-Random Number Generator
DES Data Encryption Standard
DHCP Dynamic Host Configuration Protocol
DIL/DIP Dual In-Line Package
DNS Domain Name System
DNS SEC Domain Name System Security Extensions
DSA Digital Signature Algorithm
DSS Digital Signature Standard
EAPO Eurasian Patent Organization
334
ECC Elliptic Curve Cryptography
EMAIL Electronic Mail
EPO European Patent Office
ESP Extra-Sensory Perception
EU European Union
FAR False Acceptance Rate
FCN Fully Connected Network
FFC Finite Field Cryptography
FOREX Foreign Exchange
FRR False Rejection Rate
FTP File Transfer Protocol
FTPS FTP over SSL
GCC Gulf Cooperation Council
GCCPO Gulf Cooperation Council Patent Office
HDD Hard Disk Drive
HMAC Keyed- Hash Message Authentication Code
HTTP Hypertext Transfer Protocol
HTTPS HTTP over SSL
IACR International Association for Cryptologic Research
IATUL International Association of Technological University Libraries
IDC Identity-Based Cryptography
IEEE Institute of Electrical and Electronics Engineers, Inc.
IEICE The Institute of Electronics, Information and Communication Engineers
(電子 情報通信 学会)
IETF Internet Engineering Task Force
IFC Integer Factorization Cryptography
IIPA International Intellectual Property Alliance
ILBS International Law Book Services
IM Instant Messaging
IMAP4 Internet Message Access Protocol version 4
IP Intellectual Property
IPOS Intellectual Property Office of Singapore
IPR Intellectual Property Right
IRC Internet Relay Chat
ITRC Identity Theft Resource Center
JPO Japan Patent Office ( 日 本 経済 産業 省 特許庁 )
MaC Message Authenncation Code
MCMc Malavsian Communications and Multimedia Commission
MDC Multimedia Development Corporation Sdn Bhd
335
MDeC Multimedia Development Corporation Sdn Bhd
MePKC Memorizable Public-Key Cryptography I Memorizable Public-Key Cryptosystem
MIME Multipurpose Internet Mail Extensions
MITM Man In The Middle
MoPKC Mobile Public-Key Cryptography
MTSO Mobile Telephone Switching Office
MY Malaysia
MylPO Intellectual Property Corporation of Malaysia (Perbadanan Harta Intelek Malaysia)
NBER National Bureau of Economic Research
NIST National Institute of Standards and Technology
OAPI/AIPO Organisation Africaine de la Propriete Intellectuelle
(African Intellectual Property Organization)
OSCAR Open System for CommunicAtion in Realtime
(AOL Instant Messenger Protocol for ICQ and AIM)
OTP One-Time Password
PAKE Password- Authenitcated Key Exchange
PCPIP Paris Convention for the Protection of Industrial Property
PCT Patent Cooperation Treaty
PGP Pretty Good Privacy
Ph.D. Doctor of Philosophy
PKC Public-Key Cryptography
PKC Public-Key Cryptosystem
PLT Patent Law Treaty
PNAS Proceedings of the National Academy of Sciences
POP3 Post Office Protocol version 3
P.R.C. People's Republic of China ( 中华 人民 共和国)
PRNG Pseudo-Random Number Generator
PSTN Public Switched Telephone Network
RFC Request for Comments
RFID Radio Frequency Identification
Rlogin Remote Login in UNIX Systems
RNG Random Number Generator
R.O.C. Republic of China (中華 民國)
RSA Ri vest- Sh amir- Adleman Public-Key Cryptography
S/MIME Secure I Multipurpose Internet Mail Extensions
SATA Serial Advanced Technology Attachment
SD Statutory Declaration
SFTP Secure FTP over SSH
SHA Secure Hash Algorithm
336
SHS Secure Hash Standard
SIP Session Initiation Protocol
SMTP Simple Mail Transfer Protocol
SIPO State Intellectual Property of the P.R.C. ( 中华人 民 共和 国 国 家 知 识产权 局 )
SMS Short Message Service
SNMP Simple Network Management Protocol
SPC Strasbourg Patent Convention
SPEKE Simple Password Exponential Key Exchange
SPLT Substantive Patent Law Treaty
SRP-6 Secure Remote Password Protocol version 6
SSH Secure Shell
SSL Secure Sockets Layer
TAC Transaction Authorisation Code or Transaction Authentication Code
TAP Transaction Authorization Pin
TELNET Telecommunication Network
TIPO The Intellectual Property Office of Ministry of Economic Affairs, R.O.C.
( 中 華 民 國經濟 部 智慧 財產 局)
TLS Transport Laver Security
TRIPS Agreement on Trade Related Aspects 01 intellectual Property Rights
TSIG Transaction SIGnature Protocol
TSA Timestamping Authority
TSP Time-Stamp Protocol
TTP Trusted Third Party
UI Utility Innovation
UK United Kingdom
UKCS UK Copyright Service
UN United Nations
UNESCOBKK UNESCO Bangkok
UNODC United Nations Office on Drugs and Crime
US United States
USA United States of America
USB Universal Serial Bus
USCO US Copyright Office
USCOC US Chamber of Commerce
USPTO US Patent and Trademark Office
MSVS Microsoft Visual Studio
WIPO World Intellectual Property Organization
WTO World Trade Organization
337
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11