Great Encryption Algo PDF

PERFORMANCE EVALUATION OF NEW
ENCRYPTION ALGORITHMS WITH

EMPHASIS ON PROBABILISTIC
ENCRYPTION & TIME STAMP IN NETWORK
SECURITY
Thesis Submitted to
Acharya Nagarjuna University
FOR THE AWARD OF THE DEGREE OF
DOCTOR OF PHILOSOPHY
IN
COMPUTER SCIENCE AND ENGINEERING
BY
A.V.N.KRISHNA
Under the Guidance of
Prof. A.VINAYA BABU
Department of Computer Science and Engineering

Acharya Nagarjuna University,
Nagarjuna Nagar – 522 510
Andhra Pradesh, India, Aug. ‘09
DECLARATION
I hereby declare that the entire thesis work entitled, “PERFORMANCE

EVALUATION OF NEW ENCRYPTION ALGORITHMS WITH EMPHASIS ON
PROBABILISTIC ENCRYPTION & TIME STAMP IN NETWORK SECURITY”
submitted to the department of Computer Science and Engineering, Acharya
Nagarjuna University, in fulfillment of the requirement for the award of the
degree of Doctor of Philosophy is a bonafide record of my own work carried out
under the supervision of Dr. A.VINAYA BABU, Professor of Computer
Science, Director, Admissions, JNTUH, Hyderabad. I further declare that the
thesis either in part or full, has not been submitted earlier by me or others for the
award of any degree in any University.
Date: 01 -08 - 2009
HYDERABAD
A.V.N.KRISHNA
Department of Computer Science and Engineering
Acharya Nagarjuna University, Nagarjuna Nagar – 522 510
Andhra Pradesh, INDIA.

E_mail:hari_avn@rediffmail.com Phone: 91-9849520995; Res: 91-
40-24047524
JAWAHARLAL NEHRU TECHNOLOGICAL
UNIVERSITY
HYDERABAD, A.P., INDIA
Dr. A.VINAYA BABU, B.E, M.E (ECE), M.TECH (CSE), Ph. D (CS).
Professor of Computer Science & Engineering
DIRECTOR, ADMISSIONS,
JNTUH, HYDERABAD, INDIA.
CERTIFICATE
This is to certify that the present dissertation work entitled “P

PERFORMANCE
EVALUATION OF NEW ENCRYPTION ALGORITHMS WITH EMPHASIS ON
PROBABILISTIC ENCRYPTION & TIME STAMP IN NETWORK SECURITY”
submitted by Mr. A.V.N.KRISHNA for the award of the Degree of Doctor of
Philosophy in Computer Science and Engineering to the Acharya Nagarjuna
University, Nagarjuna Nagar is a record of bonafide research work actually
carried out under my supervision in the Computer Science and Engineering
department of this University. The results embodied in this dissertation have not
been submitted for any other degree or diploma.
Date: 01- 08 - 2009
HYDERABAD Prof. A.VINAYA BABU

Acknowledgements
I owe an inestimable debt of gratitude to Professor Dr. A.Vinaya Babu, Director,

Admissions, Jawaharlal Nehru Technological University, Hyderabad, for help and
guidance in the successful completion of this dissertation piece. Professor Vinaya Babu,
as a Mentor, has been a dependable source of encouragement and guidance in and
through this continuous effort. I owe an inestimable debt of gratitude to Professor
S.N.N.Pandit, Director for Quantitative Studies, Osmania University, for his motivation
and encouragement in my pursuit of the Ph.D. degree.
Professor S.N.N.Pandit was a subtle motivator of my participation and
presentation of my work in International Conference on Information Systems,
Technology & Management( CISTM 05), that gave me self-confidence in pursuing the
research work to completion.
I am grateful and respectable to Dr.I.Ramesh Babu, Head of the Computer
Science & Engineering Dept. for the constant encouragement given in completion of this
work. Equally I am grateful to Dr.Satya Prasad, Chairman, Board of Studies, Computer
Science, for his continues monitoring of my work in its improvement to a standard level.
I am very thankful to Professor Trimurty for his valuable suggestions in improving this
work.
I am grateful to Sri K. Krishna Rao & Pradeep Reddy, Correspondent/Director,
Indur Institute of Engineering & Technology, Indur Society, for their untiring support
rendered to me in completion of my work. I am also grateful to Dr. B. Vishnu Vardhan
and Prof. P.Nageswara Rao for their continuous support in pursuing this work. I am also
grateful to Dr. L.Pratap Reddy, H.O.D, E.C.E. dept., J.N.T.U.H. and Dr. C.Raghavendra
Rao, Professor, Computer Information Systems, Hyderabad Central University,
Hyderabad, for their continuous support in pursuing this work. And also I am very much
grateful to my colleagues in Computer Science department, and Management of Indur
Institute of Engineering & Technology for their continuous encouragement in completion
of this work. Last but not least, I wish to acknowledge my Family member’s role in my
coming this far.
CONTENTS
Abstract …
I
List of Tables …
ii
List of Figures …
Vi
1 Introduction …
1 – 08
2 Literature Review …
09 – 34
2.1 Symmetric encryption schemes …

09-19
2.1.1 Block Ciphers …
10
2.1.1.1 DES Algorithm …
10
2.1.1.1.1 Electronic Code Book …
11
2.1.1.1.2 Cipher Block Chaining Mode 11
2.1.1.1.3 Cipher Feed Back Mode… 11
2.1.1.1.4 Output Feed Back Mode 12
2.1.1.2 Triple DES 12
2.1.1.3 Homomorphic DES 12
2.1.1.4 AES Algorithm 13
2.1.1.5 Dynamic Substitution 13-15
2.1.1.5.1 The Vernam Cipher 14
2.1.1.5.2 Cryptographic Combiners 14
2.1.1.5.3 Substitution Ciphers 14
2.1.1.5.4 Dynamic Substitution 15
2.1.1.5.5 Black Box Analysis 15
2.1.1.5.6 Poly Alphabetic Dynamic Substitution 15
2.1.1.5.7 Internal State 15
2.1.1.6 Nonces 16
2.1.1.6.1 Nonce Based Encrytpion 16
2.1.1.7 One Time Pad Encryption 16
2.1.2 Stream Ciphers 17
2.1.2.1 RC4 Encryption Algorithm 17
2.1.2.1.1 Algorithm Features 17
2.1.2.1.2 Algorithm Strengths 18
2.1.3 Adoptability of Different functions in Cryptography 18
2.2 Public Key Encryption 19-21
2.2.1 Key Length & Encryption Strength… 20
2.2.2 RSA Algorithm 20
2.2.3 Digital Signatures 21
2.3 Probabilistic Encryption 22-25
2.3.1 Definition of the Problem 23
2.3.1.1 Chosen Cipher Text Attack 23
2.3.1.2 Neural Networks in Cryptography 24
2.3.1.3 Knapsack Based Crypto Systems 24
2.3.1.4 On Probabilistic Encryption Schemes for Encryption Using Non Linear 25
code.
2.4 Numerical Model for data Development 25-29
2.4.1 Partial differential equations 25
2.4.2 Numerical Data Analysis 26
2.4.2.1 Discretization Methods 26
2.4.2.2 Control Volume Formulation 27
2.4.2.3 Steady one dimensional data flow 27
2.4.2.4 Grid spacing 27
2.4.2.5 Boundary Conditions 28
2.4.2.6 Solution of Linear Algebraic Equations 28
2.5 Key Distribution Mechanism 29-30
3 Methodology of the Proposed Work 31 – 85
…
.
3.1 Problem Definition …

32
3.2 Algorithm for Generating Sequence & Basins of Algorithm 1 …
33-41
3.2.1 A New Substitution Block Cipher …
42– 50
Algorithm(Model 1)
3.2.1.1 New Substitution cipher for data security …
42
3.2.1.2 Features of the Model1 …
42
3.2.1.3 Some of the advantages of the Model1 …
43
3.2.1.4 Example …
44
3.2.1.5 Computing power of the model1 …
45
3.2.1.6 Complexity of the model by its Construction & Strength 46
3.2.1.7 Avalanche Effect on the proposed Model1 48
3.2.1.8 Security Analysis 50
3.2.1.9 Conclusion 50
3.2.2 A New Variable length Key Block Cipher …
51 – 62
Algorithm(Model 2)
3.2.2.1 New variable length key Substitution cipher …
51
3.2.2.2 Features of the Model 2 …
52
3.2.2.3 Some of the advantages of the Model …
53
3.2.2.4 Example …
53
55
3.2.3 A New Probabilistic Stream Cipher Algorithm …
63 –76
(Model 3)
3.2.3.1 Definition of New Probabilistic Stream cipher …
63
3.2.3.2 Features of the Model3 …
65
3.2.3.3 Some of the advantages of the Model3 …
66
3.2.3.4 Example …
66
68
3.3 A New Mathematical Model with a key & a Time …
77– 85
stamp for encryption mechanism.
3.3.1 Mathematical modeling of the problem …

78
3.3.2 Linear Data flow problem …
78
3.3.3 Procedure for generating data from coefficients by tridiogonal method 79
3.3.4 Results …
80
3.3.5 Computing power of the model …
82
3.3.6 Complexity of the model by its Construction & Strength 82
3.3.7 Avalanche Effect on the proposed Model 84
3.3.8 Security Analysis 85
3.3.9 Conclusion 85
4 Training of the Developed Model 86-103
…
.
4.1 Training of the Algorithms using different keys & Time stamps …
86
4.2 Role of Statistical tests on Values generated by the algorithms under 100
study
5 Key Distribution Mechanism 104-106
5.1 Model for generation of Hash value for given key. 104
5.2 Example 105
5.3 Identifying any Garbled Key During Transmission 105
5.4 Conclusion 105
6 Comparative Study of Different developed …
107–122
Models
6.1 Computational power …
107
6.2 Complexity by Construction & Its Strength …
110
6.3 Avalanche Effect …
112
6.4 Differential & Linear crypto Analysis …
115
6.5 Security Analysis …
115
7 Applications of the Proposed Models …
123-126
7.1 Mobile Adhoc Networks …

123
7.2 Wireless Sensor Networks …
124
7.3 Broad Casting Applications …
125
7.4 Key Encapsulating Mechanism …
126
8 Summary/Conclusions …
127 – 135
8.1 Summary/Dissertation …
127
8.2 Conclusion 130
9 Future Work …
136
9.1 Future work …

136
References
137-140
…
Publications of Author based on

141-142
Research Work …
List of Tables
Table No. 3.2.1.1: Encryption & Decryption process by Model 1…………………...44
Table No.3.2.1.2: Representation of No. of computations, Computing power,

Complexity by construction & Complexity by strength of Model 1..……………….47
Table 3.2.1.3 Identifying the variations in cipher text by slight variations in the
key (Model 1)…...………………………………………………………………………49
Table No.3.2.2.1 Encryption & Decryption Process of Model 2.……………………54
Table No.3.2.2.2 Representation of No. of computations, Computing power,

Complexity by construction & Complexity by strength of Model 2………………...58
key (Model 2)………………………………………………………………….………...60
plaintext(Model2)………….……………………………………………………………61
Table No. 3.2.3.1 Encryption & Decryption Process of Model 3…………………...67
Table No.3.2.3.2 Representing No. of computations, computing power, complexity

by construction & strength of Model 3…………………………………………..…...71
Key (Model 3)………..………………………………………………………………….74
plain text (Model 3)……………………………………………………………………..74
Table 3.2.3.5 Identifying multiple copies of cipher text for one plain text…………75
Table No. 3.3.1 Encryption & Decryption Process by Mathematical Model ..……81
Table3.3.2 Complexity by construction & strength of Model

……………………………………….83
Table 4. 1: Relationship between Random Key considered with the Basins (Sub
Keys) generated………………………………………………………………………..93
Table 4. 2: Level of Metric between Key and Basins (Sub Keys) generated for Sign
Function applied on the product of Modified ternary vector & the random matrix
key …………………………………………………………………………………......94
Table 4.3: Level of Metric between Key and Basins (Sub Keys) generated for
different functions applied on the product of Modified ternary vector & random key
.............………………………………………………………………………………….95
Table 4.4: Possible Relationship between Plain text and Cipher Text Generated in
Probabilistic encryption Algorithm (Model 4) (Depending on the random Key
Chosen)………………………………………………………………………………....96
Table 4.5: Conversion of plain text to cipher text for a 5 character Text……..…..97
Table 4. 6: Conversion of plain text to cipher text for a 10 character Text…………97
Table 4.7: Conversion of plain text to cipher text for a 15 character Text…………97
Table 4.8: Conversion of plain text to cipher text for a 20 character Text…………98
Table 4.9: Multiple Cipher Texts for a 5 Character Plain Text …………………....98
Table 6.1: Comparative Study of Number of computations and computational

overhead of generated models…………………………………………………….…..109
Table 6.2: Comparative study of Different Generated Models in terms of Security

analysis by construction and crypto Analytical Strength……………………….….112
Table 6.3: Comparative Study of Different Generated Models by Avalanche effect,

Linear & Differential Crypto Analysis & Security Analysis…………………….....116
Table 6.4: Summary of Comparative study of some models of the proposed

algorithm in terms of number of computations, computational overhead, change in
plain text , change in key, complexity of the models by their construction and
strength and security analysis, depending on the chosen key………………………117
Table 6.5.1: Comparative study of the proposed encryption algorithms like Model 1
& Model 2 with standard algorithm like DES in terms of computational overhead,
Data overhead, complexity and security analysis…………………………………...119
Table 6.5.2: Comparative study of the proposed encryption algorithms like Model 3
& Model 4 with standard algorithm like RC4 in terms of computational overhead,
Data overhead, complexity and security analysis…………………………………...120
List of Figures
Figure 1.1: Encryption Model ………....…………………………………………........2
Figure 1.2: Symmetric Encryption Model…………………………………………….3
Figure 1.3: Public Key Encryption Model…………………………………………….4
Figure 1.4: Probabilistic Encryption Model ……….......……………….………........6

Fig 4.1: Sequential Data Representation of Algorithm 1…………………………...102
Fig 4.2: Sequential Data Representation of Algorithm 2……………...………… …103
Figure 6. 1: Pictorial Representation of Comparative study of some models of the

proposed algorithm in terms of computational overhead, change in plain text
,change in key Depending on the chosen Key……………………………. …………118
Figure 6.2: Pictorial Representation of the proposed encryption algorithms like

Model 1 & Model 2 with standard algorithm like DES in terms of computational
overhead & Data overhead. …………………...121
Figure 6.3: Pictorial Representation of the proposed encryption algorithms like

Model 3 & Model 4 with standard algorithm like RC4 in terms of computational
overhead & Data overhead. …………………...121
Chapter 1
INTRODUCTION
T
he necessicity of information security with in an organization have undergone
major changes in the past and present times. In the earlier times physical means
is used to provide security to data. With the advent of computers in every field,
the need for software tools for protecting files and other information stored on the
computer became important. The important tool designed to protect data and thwart
illegal users is computer security.
With the introduction and revolution in communications, one more change that
affected security is the introduction of distributed systems which requires carrying of data
between terminal user and a set of computers. Network security measures are needed to
protect data during their transmission. The mechanisms used to meet the requirements
like authentication and confidentiality are observed to be quite complex.
Security mechanisms usually involve more than a particular algorithm or protocol
for encryption & decryption purpose and as well as for generation of sub keys to be
mapped to plain text to generate cipher text. It means that participants be in possession of
some secret information (Key), which can be used for protecting data from unauthorized
use
rs. Thus a model has to be developed within which security services and
mechanisms can be viewed.
To identify and support the security services of an organization at its effective
level, the manager needs a systematic way. One approach is to consider three aspects of
information security that is Security attack, Security mechanism and Security services.
Security attack identifies different modes by which intruder tries to get unauthorized
information and the services are intended to counter security attacks, and they make use
of one or more security mechanisms to provide the service.
As the importance of information systems is ever growing in all most all fields,
electronic information takes on many of the roles, earlier they being done on papers. Few
information integrity functions that the security mechanism has to support are security
and confidentiality of the data to be transmitted and authentication of users.
There is no single mechanism that will provide all the services specified. But we
can identify a very important mechanism that supports all forms of information integrity
is cryptographic technique. Encryption of information is the most common means of
providing security. A model for encryption can be represented by the following Figure
1.1.
Plain text Cipher text

Sender’s converted converted Receiver’s
Message to cipher back to plain message
text using text using the
key same key
Figure 1.1 Encryption Model

This general model shows that there are four basic tasks in designing a particular
security service.
1. Designing an algorithm for performing encryption & decryption process.
2. Generating the secret information with the help of algorithm of step 1.
3. Identifying methods for the distribution and sharing of secret information.
4. Identifying rules to be used by both the participating parties to make it secured.
A crypto system is an algorithm, plus all possible plain texts, cipher texts and
keys. There are two general types of key based algorithms: symmetric and public key.
With most symmetric algorithms, the same key is used for both encryption and
decryption, as shown in Figure 1.2

Sender’s converted to converted
Message cipher text back to plain Receiver’s
using private text using the message
key same private
key
Figure 1.2 Symmetric-key encryption
The process of symmetric-key encryption can be very fast as the users do not
experience any significant time delay because of the encryption and decryption.
Symmetric-key encryption provides security to data as the key is shared only by the
participating parties. It also provides a degree of authentication, since information
encrypted with one symmetric key cannot be decrypted with any other symmetric key.
Thus, as long as the symmetric key is kept secret by the two parties using it to encrypt
communications, each party can be confident that it is communicating with the other as
long as the decrypted messages specify a meaningful sense.
Symmetric-key encryption will be successful only if the symmetric key is kept
secured by the two parties involved. If anyone else discovers the key, it affects both
confidentiality and authentication. The success of a symmetric algorithm rests in the key,
divulging the key means that any one could encrypt and decrypt messages. As long as the
communication needs to remain secure, the key must be protected between the
participating parties.
Encryption and decryption with a symmetric algorithm are denoted by

E K (M) = C
D K (M) = P
Symmetric algorithms can be divided into two categories. Some operate on the plain text
a single bit or byte at a time, these are called stream algorithms or stream ciphers. Others
operate on group of bits or characters. Such algorithms are called block algorithms.
Public Key algorithms use two keys, one key for encryption and the other for
decryption. One key can be called as public key which can be declared public and the
other one is private that is, the key is known only to the particular participating party.
And also public key cryptography can be used for digital signing as it supports
authentication of users. The information encrypted with one key will only be decrypted
with the other key. Further more the decryption key cannot be calculated from the
encryption key. Figure 1.3 shows a simplified view of the way public-key encryption
works.

Sender’s converted to converted
Message cipher text back to plain Receiver’s
using public text using the message
key of private key of
receiver receiver
Figure 1.3 Public-key encryption

Compared with symmetric-key encryption, public-key encryption requires more
computation and is therefore not always appropriate for large amounts of data. However,
it's possible to use public-key encryption to send a symmetric key, which can then be
used to encrypt additional data. This is the approach used by the SSL protocol. This
provides Authentication, Integrity & Confidentiality of Information at low computing
power. Since authentication of the users is very important in applications like e-
commerce and other similar applications, public key cryptography is of much use.
Encryption and decryption can be represented in a public key scheme is
E Kpu(M) = C
D Kpr(C) = M
Where Kpu is the public key and Kpr is the private key.
In public key encryption there is always a possibility of some information being

leaked out. A crypto analyst tries to get some information based on ones public key.
Complete information can not be gained here but a part of information may be gained. In
probabilistic Encryption, multiple cipher texts are generated for one plain text, a
cryptanalyst can not generate any information by chosen plain text and chosen cipher text
attacks.
Plain text Multiple

Sender’s converted to Cipher texts
Message multiple converted Receiver’s
cipher texts back to plain message
using key text using the
same key
Figure 1.4 Probabilistic encryption

Security Analysis of algorithms: Different algorithms offers different degrees of
security, it depends on how hard they are to break. If the cost required to break an
algorithm is greater than the value of the encrypted data, then the algorithm is supposed
to be safe. If the time required breaking an algorithm is longer than the time that the
encrypted data must remain secret, and then also it is safe. If the amount of data
encrypted with a single key is less than the amount of data necessary to break the
algorithm, it is supposed to be safe.
An algorithm is unconditionally secure if, it is difficult to recover the plain text in

spite of having substantial amount of cipher text. In such circumstances, only a one time
pad is unbreakable in a cipher text only attack, simply by trying every possible key one
by one and by checking whether the resulting plain text is meaningful. This is called a
brute force attack. Cryptography is more concerned with crypto systems that are
computationally infeasible to break. Any algorithm is considered computationally secure
if it cannot be broken with available resources.
The complexity of an attack can be measured as Data Complexity, the amount of

data needed as input to the attack, processing complexity, the time needed to perform the
attack and storage requirements which are the amount of memory needed to do the attack
which is space complexity.
As a thumb rule, the complexity of an attack is taken to be minimum of these

three factors. Another classification of complexities is by complexity of the algorithm by
its construction and complexity of the algorithm by its strength. By its construction, the
time complexity of the algorithm can be calculated by executing through the steps of the
algorithm, which will be referred as O(n). Complexities can also be expressed as orders
of magnitude. If the length of the key is k, then the processing complexity is given by 2k .
It means that 2 k operations are required to break the algorithm. Then the complexity of
the algorithm is said to be exponential in nature.
A desirable property of any encryption algorithm is that a small change in plain
text or the key should produce significant change in cipher text. Such an effect is known
as avalanche effect. The more the avalanche affects of the algorithm, the better the
security. Crypto analysis is the study of recovering the plain text with out access to the
key. It may also find weakness in a crypto system that identifies patterns which can be
useful in knowing the previous results.
An attempted crypto analysis is called an attack. There are five types of attack. Each of
them assumes that the crypto analyst has complete knowledge of the encryption
algorithm used.
1. Cipher text only attack: Here the intruder is in hold of cipher text only. The
crypto analyst has cipher text of several messages, all of which have been encrypted
using the same encryption algorithm. The crypto analyst’s job is to recover the plain text
or the key used to encrypt the messages, in order to decrypt other part of messages
encrypted with the same keys.
2. Known Plaintext attack: The crypto analyst is in possession of pairs of known
plain text and cipher text. His job is to get the key used to encrypt the messages or an
algorithm to decrypt any messages encrypted with the same key.
3. Chosen Plaintext Attack (CPA): Here the crypto analyst is in hold of not only
cipher text but also parts of chosen plain text. Here the intruder is identified to be placed
at encryption site to do the attack. Differential crypto analysis is an example of this mode.
4. Chosen cipher text attack (CCA): Under the CCA model, the crypto analyst
is in possession of chosen cipher text and corresponding plain text being decrypted from
the private key. After it has chosen the messages, however, it only has access to an
encryption machine.
5. Chosen text: In this model, the analyst posses the encipher algorithm, Cipher
text to be decrypted, chosen plain text messages and corresponding cipher texts,
fabricated cipher text with the corresponding decrypted plain texts developed by the
private key.
Present work:
In this work an attempt has been made to generate two algorithms which provide
security to data transmitted. The first algorithm considers a random matrix key which on
execution by a series of steps generates a sequence. This sequence is used a sub key to
build three different encryption models. Each model can be used for encryption of data.
The second algorithm considers not only the key but also initialization vector and a time
stamp to generate sub keys which are used for encryption process. And also a mechanism
has been discussed which identifies any garbled key while transmitted from the Key
Distribution Centre.
In this work both the algorithms are discussed in terms of computational security,
computational complexity and computational overhead. Both the algorithms are studied
for their strengths and limitations. A crypto analytical study of the algorithms with
emphasis on probabilistic encryption is also considered in this study.
The encryption algorithms are compared with standard algorithms like RC4 and
DES. The algorithms are also discussed in terms of its applications and also about their
advantages and limitations in network security environment.
Chapter 2
LITERATURE SURVEY
A crypto system [1,7,8,17] is an algorithm which include all possible plain texts,
cipher texts and keys. There are two general types of key based algorithms:
symmetric and public key.
2.1 Symmetric Encryption Schemes:

With symmetric-key encryption, the encryption key can be calculated from the
decryption key and vice versa. With most symmetric algorithms, the same key is used for
both encryption and decryption, as shown in Figure 1.1. Implementations of symmetric-
key encryption can be highly efficient, so that users do not experience any significant
time delay as a result of the encryption and decryption. Symmetric-key encryption also
provides a degree of authentication, since information encrypted with one symmetric key
cannot be decrypted with any other symmetric key. Thus, as long as the symmetric key is
kept secret by the two parties using it to encrypt communications, each party can be sure
that it is communicating with the other as long as the decrypted messages continue to
make sense.
Encryption functions normally take a fixed-size input to a fixed-size output, so
encryption of longer units of data must be done in one of two ways: either a block is
encrypted at a time and the blocks are somehow joined together to make the cipher text,
or a longer key is generated from a shorter one and XOR'd against the plaintext to make
the cipher text. Schemes of the former type are called block ciphers, and schemes of the
latter type are called stream ciphers.
2.1.1 Block ciphers

Block ciphers take as input the key and a block, often the same size as the key.
Further, the first block is often augmented by a block called the initialization vector,
which can add some randomness to the encryption.
2.1.1.1 DES Algorithm:

The most widely used encryption scheme is based on Data Encryption Standard
(DES). There are two inputs to the encryption function, the plain text to be encrypted and
the key. The plain text must be 64 bits in length and key is of 56 bits. First, the 64 bits of
plain text passes through an initial permutation that rearranges the bits. This is fallowed
by 16 rounds of same function, which involves permutation & substitution functions.
After 16 rounds of operation, the pre output is swapped at 32 bits position which is
passed through final permutation to get 64 bit cipher text.
Initially the key is passed through a permutation function. Then for each of the 16
rounds, a sub key is generated by a combination of left circular shift and permutation.
At each round of operation, the plain text is divided to two 32 bit halves, and the
fallowing operations are executed on 32 bit right halve of plain text. First it is expanded
to 48 bits using a expansion table, then X-ORed with key, then processed in substitution
tables to generate 32 bit output. This output is permuted using predefined table and
XORed with left 32 bit plain text to form right 32 bit pre cipher text of first round. The
right 32 bit plain text will form left 32 bit pre cipher text of first round.
Decryption uses the same algorithm as encryption, expect that the application of
sub keys is reversed. A desirable property of any encryption algorithm is that a small
change in either plain text or the key should produce a significant change in the cipher
text. This effect is known as Avalanche effect which is very strong in DES algorithm.
Since DES is a 56 bit key encryption algorithm, if we proceed by brute force attack, the
number of keys that are required to break the algorithm is 2 56 . But by differential crypto
47
analysis, it has been proved that the key can be broken in 2 combinations of known
41
plain texts. By linear crypto analysis it has been proved that, it could be broken by 2
combinations of plain text.
The DES algorithm is a basic building block for providing data security. To apply
DES in a variety of applications, four modes of operations have been defined. These four
models are intended to cover all possible applications of encryption for which DES could
be used. They involve using a initialization vector being used along with key to provided
different cipher text blocks.
2.1.1.1.1 Electronic Code Book (ECB) mode: ECB mode divides the plaintext
into blocks m1, m2, ..., mn, and computes the cipher text ci = Ei(mi). This mode is
vulnerable to many attacks and is not recommended for use in any protocols. Chief
among its defects is its vulnerability to splicing attacks, in which encrypted blocks from
one message are replaced with encrypted blocks from another.
2.1.1.1.2 Cipher Block Chaining (CBC) mode: CBC mode remedies some of
the problems of ECB mode by using an initialization vector and chaining the input of one
encryption into the next. CBC mode starts with an initialization vector iv and XORs a
value with the plaintext that is the input to each encryption. So, c1 = Ek(iv XOR m1) and
ci = Ek(ci-1 XOR mi). If a unique iv is used, then no splicing attacks can be performed,
since each block depends on all previous blocks along with the initialization vector. The
iv is a good example of a nonce that needs to satisfy Uniqueness but not
Unpredictability.
2.1.1.1.3 Cipher Feed-Back (CFB) mode: CFB mode moves the XOR of CBC
mode to the output of the encryption. In other words, the cipher text c1 = p1 XOR
Sj(E(IV)). This mode then suffers from failures of Non-Malleability, at least locally to
every block, but changes to ciphertext do not propagate very far, since each block of
ciphertext is used independently to XOR against a given block to get the plaintext.
These failures can be seen in the following example, in which a message m = m1
m2 ... mn is divided into n blocks, and encrypted with an iv under CFB mode to c1 c2 ...
cn. Suppose an adversary substitutes c'2 for c2. Then, in decryption, m1 = Ek(iv) XOR
c1, which is correct, but m'2 = Ek(c1) XOR c'2, which means that m'2 = m2 XOR c2
XOR c'2, since m2 = Ek(c1) XOR c2. Thus, in m2, the adversary can flip any bits of its
choice. Then m'3 = Ek(c'2) XOR c3, which should lead to random looking message not
under the adversary's control, since the encryption of c'2 should look random. But m4 =
Ek(c3) XOR c4 and thereafter the decryption is correct.
2.1.1.1.4 Output Feed-Back (OFB) mode OFB mode modifies CFB mode to
feed back the output of the encryption function to the encryption function without XOR-
ing the cipher text.
2.1.1.2 Triple DES:

Given the potential vulnerability of DES to brute force attack, a new mechanism
is adopted which uses multiple encryptions with DES and multiple keys. The simplest
form of multiple encryptions has two encryption stages and two keys. The limitation with
this mechanism is it is susceptible to meet in the middle attack. An obvious counter to
meet in the middle attack and reducing the cost of increasing the key length, a triple
encryption method is used, which considers only two keys with encryption with the first
key, decryption with the second key and fallowed by encryption with the first key. Triple
DES is a relatively popular alternative to DES and has been adopted for use in key
management standards.
2.1.1.3 Homomorphic DES:

A variant of DES called a homophonic DES [9] is considered. The DES algorithm
is strengthened by adding some random bits into the plaintext, which are placed in
particular positions to maximize diffusion, and to resist differential attack. Differential
attack makes use of the exclusive-or homophonic DES. In this new scheme, some
random estimated bits are added to the plaintext. This increases the certain plaintext
difference with respect to the cipher text.
A homophonic DES is a variant of DES that map search plaintext to one of many
cipher texts (for a given key). In homophonic DES a desired difference pattern with the
cipher text will be suggested with some key values including the correct one, oppositely
wrong pairs of cipher text. For a difference pattern which 56-bit plaintext to a 64-bit
cipher text using a 56-bit key. In this scheme, eight random bits are placed in specific
positions of the 64-bit input data block to maximize diffusion.
For example, the random bits in HDESS are the bit- positions 25, 27, 29, 31, 57,
59, 61 and 63. In this algorithm, after the initial permutation and expansion permutation
in the first round, these eight random bits will spread to bits 2, 6, 8, 12, 14, 18, 20, 24, 26,
30, 32, 36, 38,42,44,48 of the 48-bit input block to the S-boxes and will affect the output
of all the S-boxes. The 48 expanded bits must be exclusive-or’d with some key before
proceeding to the S-boxes, thus two input bits into the S-boxes derived from the same
random bit may have different values. This says that the random bits do not regularize the
input to the S-boxes, that is, the property of confusion does not reduce while we try to
maximize diffusion.
The decryption of the homophonic DES is similar to the decryption of DES. The
only difference is that eight random bits must be removed to get the original plaintext (56
bits). A homophonic DES can easily be transformed into a triple-encryption version by
concatenating a DES decryption and a DES encryption after the homophonic DES.
Security analysis: Thus there is a probability of 1/256 between a pair of texts. The
differential crypto analysis is also difficult on this mechanism. The diffusion of bits is
also more in this mode. Thus this mechanism provides some probabilistic features to DES
algorithm which makes it stronger from differential and linear crypto analysis.
2.1.1.4 AES:
The Advanced Encryption Standard (AES) was chosen in 2001. AES is also an
iterated block cipher, with 10, 12, or 14 rounds for key sizes 128, 192, and 256 bits,
respectively. AES provides high performance symmetric key encryption and decryption.
2.1.1.5 Dynamic substitution:

An apparently new cryptographic mechanism [39] which can be described as
dynamic substitution is discussed in the fallowing topic. Although structurally similar to
simple substitution, dynamic substitution has a second data input which acts to re-arrange
the contents of the substitution table. The mechanism combines two data sources into a
complex result; under appropriate conditions, a related inverse mechanism can then
extract one of the data sources from the result. A dynamic substitution combiner can
directly replace the exclusive-OR combiner used in Vernam stream ciphers. The various
techniques used in Vernam ciphers can also be applied to dynamic substitution; any
cryptographic advantage is thus due to the additional strength of the new combiner.
2.1.1.5.1 The Vernam Cipher: A Vernam cipher maps plaintext data with a
pseudo-random sequence to generate cipher text. Since each ciphertext element from a
Vernam combiner is the (mod 2) sum of two unknown values, the plaintext data is
supposed to be safe. But this mode is susceptive to several cryptanalytic attacks,
including known plain text and cipher text attacks. And if the confusion sequence can be
penetrated and reproduced, the cipher is broken. Similarly, if the same confusion
sequence is ever re-used, and the overlap identified, it becomes simple to break that
section of the cipher.
2.1.1.5.2 Cryptographic Combiners: An alternate approach to the design of a

secure stream cipher is to seek combining functions which can resist attack; such
functions would act to hide the pseudo-random sequence from analysis.
The mechanism of this work is a new combining function which extends the weak
classical concept of simple substitution into a stronger form suitable for computer
cryptography.
2.1.1.5.3 Substitution Ciphers: In simple substitution ciphers each plain text

character is replaced with fixed cipher text character. But this mechanism is weak from
statistical analysis methods where by considering the rules of the language, the cipher can
be broken. This work is concerned with the cryptographic strengthening of the
fundamental substitution operation through dynamic changes to a substitution table. The
substitution table can be represented as a function of not only input data but also a
random sequence. This combination gives a cryptographic combining function; such a
function may be used to combine plaintext data with a pseudo-random sequence to
generate enciphered data.
2.1.1.5.4 Dynamic Substitution: A simple substitution table supported with

combining function gives the idea of dynamic substitution. A substitution table is used to
translate each data value into an enciphered value. But after each substitution, the table is
re-ordered. At a minimum, it makes sense to exchange the just-used substitution value
with some entry in the table selected at random. This generally changes the just-used
substitution value to help prevent analysis, and yet retains the existence of an inverse, so
that the cipher can be deciphered.
2.1.1.5.5 Black Box Analysis: Dynamic substitution may be considered to be a

black box, with two input ports Data In and Random In, and one output port Combiner
Out. In the simple version, each data path has similar width; evidently the mechanism
inside the box in some way combines the two input streams to produce the output stream.
It seems reasonable to analyze the output statistically, for various input streams.
2.1.1.5.6 Polyalphabetic Dynamic Substitution: A means to defend to known-

plaintext and chosen-plaintext attacks would be to use multiple different dynamic
substitution maps and to select between them using a hidden pseudo-random sequence.
Thus the dynamic substitution if free from statistical attacks where each character of
plain text is replaced with multiple characters of cipher text which makes the mechanism
robust.
2.1.1.5.7 Internal State: Dynamic substitution contains internal data which after
initialization is continuously re-ordered as a consequence of both incoming data streams;
thus, the internal state is a function of initialization and all subsequent data and confusion
values. The changing internal state of dynamic substitution provides necessary security to
the data streams.
Thus dynamic substitution provides a probabilistic nature to the enciphering

mechanism. The limitation with this scheme is, not only different dynamic substitution
tables has to be maintained but also the pseudo random sequence which selects between
these dynamic substitution tables has to be shared between sender and receiver.
2.1.1.6 Nonces
A nonce [34] is a bit string that satisfies Uniqueness, which means that it has not
occurred before in a given run of a protocol. Nonces might also satisfy Unpredictability,
which effectively requires pseudo-randomness: no adversary can predict the next nonce
that will be chosen by any principal. There are several common sources of nonces like
counters, time slots and so on.
2.1.1.6.1 Nonce Based Encryption: In this work a different formalization for

symmetric encryption is envisaged. The encryption algorithm is made to be a
deterministic function, but it is supported with initialization vector (IV). Efficiency of
the user is made success of this mode. The IV is a nonce like value, used at most once
within a session. Since it is used at most once having any sort of crypto analysis is
practically not possible which provides sufficient security.
2.1.1.7 One-Time Pad Encryption

One more encryption mechanism for providing security to data is one time pad
[17] encryption. The functions are computed as follows: A and B agree on a random
number k that is as long as the message they later want to send.
Ek(x) = x XOR k
Dk(x) = x XOR k
Note that since k is chosen at random and not known to an adversary, the output of this
scheme is indistinguishable to an adversary from a random number. But it suffers from
several limitations. It is susceptible to chosen plain text and chosen cipher text attacks.
Again the limitation is here is sharing of one time keys by the participating parties of the
encryption scheme. As a new key is always used for encryption, a continuous sharing of
key mechanism has to be employed by the participating parties.
2.1.2 Stream ciphers
Unlike block ciphers, stream ciphers [18] (such as RC4) produce a pseudo-
random sequence of bits that are then combined with the message to give an encryption.
Since the combining operation is often XOR, naive implementations of these schemes
can be vulnerable to the sort of bit-flipping attacks on Non-Malleability. Two types of
stream ciphers exist: synchronous, in which state is kept by the encryption algorithm but
is not correlated with the plaintext or cipher text, and self synchronizing, in which some
information from the plaintext or cipher text is used to inform the operation of the cipher.
2.1.2.1 RC4 Encryption Algorithm:
Ronald Rivest of RSA developed the RC4 algorithm, which is a shared key
stream cipher algorithm requiring a secure exchange of a shared key. The algorithm is
used identically for encryption and decryption as the data stream is simply XORed with
the generated key sequence. The algorithm is serial as it requires successive exchanges of
state entries based on the key sequence. Hence implementations can be very
computationally intensive. In the algorithm the key stream is completely independent of
the plaintext used. An 8 * 8 S-Box (S0 S255), where each of the entries is a permutation
of the numbers 0 to 255, and the permutation is a function of the variable length key.
There are two counters i, and j, both initialized to 0 used in the algorithm.
2.1.2.1.1 Algorithm Features: 1.It uses a variable length key from 1 to 256
bytes to initialize a 256-byte state table. The state table is used for subsequent generation
of pseudo-random bytes and then to generate a pseudo-random stream which is XORed
with the plaintext to give the cipher text. Each element in the state table is swapped at
least once.
2. The key is often limited to 40 bits, because of export restrictions but it is
sometimes used as a 128 bit key. It has the capability of using keys between 1 and 2048
bits. RC4 is used in many commercial software packages such as Lotus Notes and Oracle
Secure.
3. The algorithm works in two phases, key setup and ciphering. During a N-bit
key setup (N being your key length), the encryption key is used to generate an encrypting
variable using two arrays, state and key, and N-number of mixing operations. These
mixing operations consist of swapping bytes, modulo operations, and other formulas.
2.1.2.1.2 Algorithm Strengths: The difficulty of knowing which location in the
table is used to select each value in the sequence. A particular RC4 Algorithm key can be
used only once and Encryption is about 10 times faster than DES. Algorithm Weakness:
One in every 256 keys can be a weak key. These keys are identified by cryptanalysis that
is able to find circumstances under which one of more generated bytes are strongly
correlated with a few bytes of the key.
Thus some symmetric encryption algorithms have been discussed in this chapter.
They varies from block ciphers like DES, Triple DES, Homomorphic DES to stream
ciphers like RC4. To the symmetric encryption mechanisms concepts like application of
Nounce and dynamic substitution are discussed which provides randomness to the
encryption mechanism. This probabilistic nature to the encryption mechanism provides
sufficient strength to the algorithms against Chosen Cipher text attacks(CCA). The
security with all these mechanisms lies with proper sharing of keys among the different
participating parties.
2.1.3 Adoptability of some mathematical functions in Cryptography:

Sign Function: [31,32] This function when applied on when applied on a matrix
of values, converts all the positive values to 1, negative values to -1 & zero with 0. The
advantage of using this function in cryptography is it cannot be a reversible process ie we
cannot get back to the original matrix by applying a reverse process.
Modular Arithmetic: One more function that is widely used in cryptography is
modular arithmetic of a number with a base value. It will generate the remainder of a
number with respect to the base value. This function is widely used in public key
cryptography.
2.2 Public-Key Encryption
The most commonly used implementations of public-key [17,18] encryption are
based on algorithms patented by RSA Data Security. Therefore, this section describes the
RSA approach to public-key encryption.
Public-key encryption (also called asymmetric encryption) involves a pair of keys a
public key and a private key, used for security & authentication of data. Each public key
is published, and the corresponding private key is kept secret. Data encrypted with one
key can be decrypted only with other key.
The scheme shown in Figure 1.2 says public key is distributed and encryption
being done using this key. In general, to send encrypted data, one encrypt’s the data with
the receiver’s public key, and the person receiving the encrypted data decrypts it with his
private key.
Compared with symmetric-key encryption, public-key encryption requires more

computation and is therefore not always appropriate for large amounts of data. However,
a combination of symmetric & Asymmetric schemes can be used in real time
environment. This is the approach used by the SSL protocol.
As it happens, the reverse of the scheme shown in Figure 1.2 also works: data
encrypted with one’s private key can be decrypted only with his public key. This may not
be an interesting way to encrypt important data, however, because it means that anyone
with receiver’s public key, which is by definition published, could decipher the data. And
also the important requirement with data transfer is authentication of data which is
supported with Asymmetric encryption schemes, which is an important requirement for
electronic commerce and other commercial applications of cryptography.
2.2.1 Key Length and Encryption Strength:
In general, the strength of encryption algorithm depends on difficulty in getting

the key, which in turn depends on both the cipher used and the length of the key. For the
RSA cipher, the strength depends on the difficulty of factoring large numbers, which is a
well-known mathematical problem. Encryption strength is often described in terms of the
length of the keys used to perform the encryption, means the more the length of the key,
the more the strength. Key length is measured in bits. For example, a RC4 symmetric-key
cipher with key length of 128 bits supported by SSL provide significantly better
cryptographic protection than 40-bit keys for use with the same cipher. It means 128-bit
RC4 encryption is 3 x 1026 times stronger than 40-bit RC4 encryption. Different
encryption algorithms require variable key lengths to achieve the same level of
encryption strength.
Other ciphers, such as those used for symmetric key encryption, can use all
possible values for a key of a given length, rather than a subset of those values. Thus a
128-bit key for use with a symmetric-key encryption cipher would provide stronger
encryption than a 128-bit key for use with the RSA public-key encryption cipher.
This says that a symmetric encryption algorithm with a key length of 56 bits achieve a
equal security to Asymmetric encryption algorithm with a key length of 512 bits,
2.2.2 RSA Key Generation Algorithm
1. Two large prime numbers are considered. Let them be p,q.

2. Calculate n = pq and (φ) phi = (p-1)(q-1).
3. Select e, such that 1 < e < phi and gcd(e, phi) = 1.
4. Calculate d, the private key, such that de =1 mod phi.
One key is (n, e) and the other key is (n, d). The values of p, q, and phi should also be
kept secret.
• n is known as the modulus.

• e is known as the public key.
• d is known as the secret key.
Encryption
Sender A does the following:-
1. Get the recipient B's public key (n, e).

2. Identify the plaintext message as a positive integer m.
3. Calculate the ciphertext c = m^e mod n.
4. Transmits the ciphertext c to receiver B.
Decryption
Recipient B does the following:-
1. Consider his own private key (n, d) to compute the plain text m = c^d mod n.
2. Convert the integer to plain text form.
2.2.3 Digital signing

Sender A does the following:-
This concept can also be used in digital signing as well. The message to be transmitted is
converted to some message digest form. This message digest is converted to encryption
form using his private key. This encrypted message digest is transmitted to receiver.
Signature verification
Recipient B does the following:-
1. Using the sender’s public key, the received message digest is decrypted. From the
received message, the receiver independently computes the message digest of the
information that has been signed.
2. If both message digests are identical, the signature is valid.
Compared with symmetric-key encryption, public-key encryption provides authentication

& security to the data transmitted but requires more computation and is therefore not
always appropriate for large amounts of data.
2.3. Probabilistic encryption schemes

leaked out. Because a crypto analyst can always encrypt random messages with a public
key, he can get some information. Not a whole of information is to be gained here, but
there are potential problems with allowing a crypto analyst to encrypt random messages
with public key. Some information is leaked out every time to the crypto analyst, he
encrypts a message.
With probabilistic encryption algorithms [8,13], a crypto analyst can no longer

encrypt random plain texts looking for correct cipher text. Since multiple cipher texts will
be developed for one plain text, even if he decrypts the message to plain text, he does not
know how far he had guessed the message correctly. To illustrate, assume a crypto
analyst has a certain cipher text ci. Even if he guesses message correctly, when he
encrypts message the result will be completely different cj. He cannot compare ci and cj
and so cannot know that he has guessed the message correctly. Under this scheme,
different cipher texts will be formed for one plain text. Also the cipher text will always be
larger than plain text. This develops the concept of multiple cipher texts for one plain
text. This concept makes crypto analysis difficult to apply on plain text and cipher text
pairs.
An encryption scheme consists of three algorithms: The encryption algorithm

transforms plaintexts into cipher texts while the decryption algorithm converts cipher
texts back into plaintexts. A third algorithm, called the key generator, creates pairs of
keys: an encryption key, input to the encryption algorithm, and a related decryption key
needed to decrypt. The encryption key relates encryptions to the decryption key. The key
generator is considered to be a probabilistic algorithm, which prevents an adversary from
simply running the key generator to get the decryption key for an intercepted message.
The following concept is crucial to probabilistic cryptography:
2.3.1 Definition [Probabilistic Algorithm]:

A probabilistic algorithm [13] is an algorithm with an additional command
RANDOM that returns “0” or “1”, each with probability 1/2. In the literature, these
random choices are often referred to as coin flips.
2.3.1.1 Chosen Cipher Text Attack:

In the simplest attack model, known as Chosen Plaintext Attack (CPA) [7], the
adversary has access to a machine that will perform arbitrary encryptions but will not
reveal the shared key. This machine corresponds intuitively to being able to see many
encryptions of many messages before trying to decrypt a new message. In this case,
Semantic Security requires that it be computationally hard for any adversary to
distinguish an encryption Ek(m) from Ek(m') for two arbitrarily chosen messages m and
m'. Distinguishing these encryptions should be hard even if the adversary can request
encryptions of arbitrary messages. Note that this property cannot be satisfied if the
encryption function is deterministic! In this case, the adversary can simply request an
encryption of m and an encryption of m' and compare them. This is a point that one
should all remember when implementing systems: encrypting under a deterministic
function with no randomness in the input does not provide Semantic Security. One more
crypto analytical model is Chosen Cipher text Attack (CCA) Model. Under the CCA
model, an adversary has access to an encryption and a decryption machine and must
perform the same task of distinguishing encryptions of two messages of its choice. First,
the adversary is allowed to interact with the encryption and decryption services and
choose the pair of messages. After it has chosen the messages, however, it only has
access to an encryption machine. An advancement to CCA Model is Chosen Cipher text
Attack 2 (CCA2). CCA2 security has the same model as CCA security, except that the
adversary retains access to the decryption machine after choosing the two messages. To
keep this property from being trivially violated, we require that the adversary not be able
to decrypt the cipher text it is given to analyze.
To make these concepts of CCA & CCA2 adoptable in real time environment,
recently Canetti, Krawczyk and Nielsen defined the notion of replayable adaptive chosen
ciphertext attack [5] secure encryption. Essentially a cryptosystem that is RCCA secure
has full CCA2 security except for the little detail that it may be possible to modify a
ciphertext into another ciphertext containing the same plaintext. This provides the
possibility of perfectly replayable RCCA secure encryption. By this, we mean that
anybody can convert a ciphertext y with plaintext m into a different ciphertext y that is
distributed identically to a fresh encryption of m. It propose such a rerandomizable
cryptosystem, which is secure against semi-generic adversaries. To improve the
efficiency of the algorithm, a probabilistic trapdoor one way function is presented. This
adds randomness to the proposed work which makes crypto analysis difficult.
2.3.1.2 Neural networks in cryptography:

One more technique that is used in probabilistic encryption is to adopt Neural
Networks [14] on encryption mechanisms. Neural network techniques are added to
probabilistic encryption to make cipher text stronger. In addition to security it can also be
seen that data over head could be avoided in the conversion process A new probabilistic
symmetric probabilistic encryption scheme based on chaotic attractors of neural networks
can be considered. The scheme is based on chaotic properties of the Over storaged
Hopfield Neural Network (OHNN). The approach bridges the relationship between
neural network and cryptography. However, there are some problems in the scheme: (1)
exhaustive search is needed to find all the attractors; (2) problem exists on creating the
synaptic weight matrix.
2.3.1.3 Knapsack-based crypto systems:
Knapsack-based cryptosystems [2] had been viewed as the most attractive and the
most promising asymmetric cryptographic algorithms for a long time due to their NP-
completeness nature and high speed in encryption/decryption. Unfortunately, most of
them are broken for the low-density feature of the underlying knapsack problems. To
improve the performance of the model a new easy compact knapsack problem and
propose a novel knapsack-based probabilistic public-key cryptosystem in which the
cipher-text is non-linear with the plaintext.
2.3.1.4 On Probabilistic Scheme for Encryption Using Nonlinear Codes Mapped

from Z_4 Linear Codes:
Probabilistic encryption becomes more and more important since its ability to
against chosen-cipher text attack. To convert any deterministic encryption scheme into a
probabilistic encryption scheme, a randomized media is needed to apply on the message
and carry the message over as an randomized input [27,28]. Thus nonlinear codes
obtained by certain mapping from linear error-correcting codes are considered to serve as
such carrying media.
Thus some algorithms are discussed in literature which is symmetric and
probabilistic in nature.
2.4 Numerical Model for data development

2.4.1 Partial differential equations: Partial differential equations to model multiscale
phenomena are ubiquitous in industrial applications and their numerical solution is an
outstanding challenge within the field of scientific computing [38]. The approach is to
process the mathematical model at the level of the equations, before discretization, either
removing non-essential small scales when possible, or exploiting special features of the
small scales such as self-similarity or scale separation to formulate more tractable
computational problems. Types of data ,
1.Static: Each data item is considered free from any time based and the inferences that
can be derived from this data are also free of any time based aspects
2.Sequence. In this category of data, though there may not be any explicit reference to
time, there exists a sort of qualitative time based relationship among data values.
3.Time stamped. Here we can not only say that a transaction occurred before another but
also the exact temporal distance between the data elements. Also with the activities being
uniformly spaced on the time parameter.
4.Fully Temporal: In this category, the validity of the data elements is time dependent.
The inferences are necessarily time dependent in such cases.
2.4.2 Numerical Data Analysis
The following are the steps to generate a numerical method for data analysis[36,39].
2.4.2.1 Discretisation Methods.
The numerical solution of data flow and other related process can begin when the laws
governing these processes are represented in differential equations. The individual
differential equations follow a certain conservation principle. Each equation employs a
certain quantity as its dependent variable and implies that there must be a balance
among various factors that influence the variable.
The numerical solution of a differential equation consists of a set of numbers from
which the distribution of the dependent variable can be constructed. It means a
numerical method is equal to a experiment in which a set of experimental values gives a
means of the measured quantity in the domain under study.
Let us suppose that we decide to represent the variation of ∅ by a polynomial in x

∅ = a 0 + a 1 x + a 2 x2 + …………………..a n x n
and employ a numerical method to find the finite number of coefficients a1 ,
a2……….an. This will enable us to evaluate ∅, at any location x by substituting the
value of x and the values of a’s in the above equation.
Thus a numerical method treats as its basic unknowns the values of the dependent
variable at a finite number of location called the grid points in the calculation domain.
This method includes the task of providing a set of algebraic equations for these
unknowns and of prescribing an algorithm for solving the equations.
A discretisation equation is an algebraic equation connecting the values of ∅ for a set
of grid points. Such an equation is derived from the differential equation governing ∅
and thus expresses the same physical information as the differential information. That is
only a few grid points are represented in the given differential equation. The value of ∅
at a grid point is represented by values at its neighborhood values. As more and more
grid points are considered, the solutions of discritization equations reach the exact
solution of the corresponding differential equations.
2.4.2.2 Control Volume Formulation.
The considered area is divided into a number of grid points each with control volumes
surrounding each grid point. The differential equation is integrated over each control
volume piecewise to identify the data values. The feature of the control volume
formulation is that the output data to the control volume is equal to input data values of
the control volume. It means that conservation principle is identified over the control
volume. This characteristic exists for any number of grid points. Thus even the course
grid solution exhibits exact integral balances.
2.4.2.3 Steady One Dimensional data flow.

Steady state one-dimensional equation is given by ∂./∂x(k. ∂T/∂x) +s =0. 0 (eq. 1)
where k & s are constants. To derive the discretisation equation we shall employ the
grid point cluster. We focus attention on grid point P, which has grid points E, W as
neighbors. For one dimensional problem under consideration we shall assume a unit
thickness in y and z directions. Thus the volume of control volume is delx*1*1.
Thus if we integrate the above equation over the control volume, we get
( K ∂.T/∂X)e – (K ∂T/∂X)w + ∫S ∂X = 0.0 (eq. 2)
If we evaluate the derivatives . ∂T/ ∂X in the above equation from piece wise linear
profile , the resulting equation will be
Ke( Te – Tp)/( ∂X)e – Kw(Tp – Tw)/( ∂X)w + S *del x=0.0 where S is average value of s
over control volume. (eq. 3)
This leads to discretisation equation

apTp = aeTe + awTw +b Where ae= Ke/∂Xe (eq. 4)
aw = Kw/dXw
ap= ae+aw-sp.delX
b=se.delX .
2.4.2.4 Grid Spacing

For the grid points the distances (dX)e and (dX)w may be or may not be equal. For
simplicity we assume the grid spacing as equal on the left side and right side of grid
points. Indeed, the use of non uniform grid spacing is often desirable, for it enables us
to deploy more efficiently. Infact we shall obtain an accurate solution only when the
grid is sufficiently fine. But there is no need to employ a fine grid in regions where the
dependent variable T changes slowly with X. On the other hand, a fine grid is required
where the T_X variation is steep. The number of grid points and the way they are
distributed gives the nature of problem to be solved. Theoretical calculations using only
a few grid points specify a convenient way of learning.
2.4.2.5 Boundary Conditions

There is one grid point on each of the two boundaries. The other grid points are called
internal points, around each of which a control volume is considered. Based on the grid
points at boundary, internal grid points are evaluated by Tri diagonal matrix algorithm.
2.4.2.6 Solution Of Linear Algebraic Equations

The solution of the discretisation equations for the one-dimensional situation can be
obtained by the standard Gaussian elimination method. Because of the particularly
simple form of equations, the elimination process leads to a delightfully convenient
algorithm. For convenience in presenting the algorithm, it is necessary to use somewhat
different nomenclature. Suppose the grid points are numbered 1,2,3…ni where 1 and ni
denoting boundary points.
The discretisation equation based on equations (1-4) can be written as

Ai Ti + BiTi+1 +CiT i-1 = Di (eq. 5)
For I = 1,2,3………….ni. Thus the data value T is related to neighboring data values T
i+1 and T i-1. For the given problem
C1=0 and Bn=0;
These conditions imply that T1 is known in terms of T2. The equation for I=2, is a
relation between T1, T2 & T3. But since T1 can be expressed in terms of T2 , this
relation reduces to a relation between T2 and T3. This process of substitution can be
continued until Tn-1 can be formally expressed as Tn. But since Tn is known we can
obtain Tn-1.This enables us to begin back substitution process in which Tn-2,Tn-
3………….T3,T2 can be obtained.
For this tridiogonal system , it is easy to modify the Gaussian elimination procedures to
take advantage of zeros in the matrix of coefficients.
Referring to the tridiogonal matrix of coefficients above, the system is put into a upper
triangular form by computing new Ai.
Ai = Ai – (C i-1 /Ai)* Bi where i = 2,3……………ni. (eq. 6)
Di= Di – (C i-1 /Ai) * Di (eq. 7)
Then computing the unknowns from back substitution
Tn = Dn / An. (eq. 8)
Then Tn = Dk – Ak * T k+1 / Ak, k= ni-1, ni-2…3,2,1. (eq. 9)
Thus a Sequence of values are generated using tridiogonal matrix algorithm which can
be used as sub key in cryptographic techniques.
2.5 Key Distribution Mechanism

In most of the schemes, a key distribution centre (KDC) is employed which handles the
task of key distribution for the participating parties. Generally two mechanisms are
employed [ 5,10,19].
In the first mechanism user A, requests KDC for a session with another user say, B.
Initially the KDC sends session key encrypted with private key of A, to the user A. This
encrypted session key is appended with encrypted session key by private key of B. On
receiving this User A, gets session key and encrypted message with private key of B.
This encrypted message is sent to B, where B decrypts it and gets the session key. Now
both A & B are in hold of session key which they can use for secured transmission of
data. Other wise it is the KDC which sends encrypted session key to the participating
parties based on the request of user.
In the second mechanism, the scenario assumes that each user shares a unique master key
with the key distribution centre. In such a case, the session key is encrypted with the
master key and sent to participating parties.
A more flexible scheme, referred to as the control vector [10]. In this scheme, each
session key has an associated control vector consisting of a number of fields that specify
the uses and restrictions for that session key. The length of the control vector may vary.
As a first step, the control vector is passed through a hash function that produces a value
which is equal to encryption key length. The hash value is XOR ed with the master key to
produce an output that is used as key to encrypt the session key. When the session key is
delivered to the user the control vector is delivered in its plain form. The session key can
be recovered only by using both master key that the user shares with the KDC and the
control vector. Thus the linkage between session key & control vector is maintained.
Some times keys get garbled in transmission. Since a garbled key can mean mega bytes
of unacceptable cipher text, this sis a problem. All keys should be transmitted with some
kind of error detection and correction bits. This is one way errors of key can be easily
detected and if required the key can be reset.
One of the most widely used methods is to encrypt a constant value with the key and to
send the first 2 to 4 bytes of that cipher text along with the key. At the receiving end, the
same thing is being done. If the encrypted constants match then the key has been
16 t
transmitted with out error. The chance of undetected error ranges from one in 2 o one
in 2 32. The limitation with this approach is in addition to the key, even the constant has
to be transmitted to participating parties.
Some times the receiver wants to check if a particular key he has, is the correct
decryption key. The naïve approach is to attach a verification block, a known header to
the plain text message before encryption. At the receiver’s side, the receiver decrypts the
header and verifies that it is correct. This works, but it gives intruder a known plain text
to help crypto analyze the system.
Chapter 3
METHODOLOGY OF THE PROPOSED WORK
Introduction
The following sequence of steps identifies the methodology adopted in this work.
1. Definition of the problem.
2. Algorithms for generation of subkeys.
3. An algorithm 1 which multiplies ternary vector and a random matrix key
to generate a sequence. Dividing the sequence generated in algorithm 1
into basins based on equality of values. Mapping of the sequence or basins
to the plain text to generate cipher text. This mapping develops 3 models
which are discussed in detail as subunits. The developed algorithm is
trained to find an optimal key.
4. An algorithm 2 which considers a key, a time stamp and an Initialization
vector to generate sub keys which are mapped to plain text to generate
cipher text.
5. Training of the developed algorithms with different keys.
6. Adopting a suitable mechanism to identify any garbled keys while
transmission from the Key distribution centre.
7. Comparative study of the developed algorithms in terms of Computing
power, their Complexity in terms construction & strength, Avalanche
effect & Security analysis.
8. Comparative study of the algorithms with standard models like DES &
RC4.
9. Summary & Conclusion of the work.
3.1 Problem Definition
1. The type of operations used for transforming plain text to cipher text. All
encryption algorithms are based on two general principles. Substitution in which
each element in the plain text is mapped to another element and transposition in
which the elements in the plain text are re arranged. Most systems involve
multiple stages of substitution and transpositions.
2. The number of keys used. If the sender and receiver use the same key, the
system is referred as symmetric, single key, secret key or conventional
encryption. If the sender and the receiver each uses a different key, the system is
referred to as asymmetric, two key or public key encryption.
3. The way in which plain text is processed. A block cipher processes the input
one block of elements at a time, producing an output block for each input block.
A stream cipher processes the input element continuously, producing output one
element at a time, as it goes along.
Present work: In the present work two algorithms are developed. The first algorithm
uses a matrix key which on multiplication with a ternary vector and applying a sign
function on the product generates a sequence. This sequence will be used to generate
three different models of substitution technique. Thus the algorithm is considered to be a
substitution algorithm which uses a single key to be shared by both the sender and
receiver, and the cipher processes the input element continuously, producing output one
element at a time. The new encryption algorithm is based on the concept of Poly
alphabet cipher which is an improvement over mono alphabet. Three models are
developed from the given algorithm [22, 23]. The first two models like Model 1 & Model
2 can be classified under block ciphers and Model 3 is a stream ciphers. Each model is
having its own advantages and limitations.
The second algorithm considers not only key but also initialization vector and a time
stamp to generate sub keys which are used for encryption process.
3.2 Algorithm for generating the sequence
1. Consider the sequence for 0 to n values where n is a positive integer.
2. Convert each element of the sequence into ternary form of a given digit
number.
3. Represent the values of step 2 in a matrix form of (n+1) * (digit number).
4. Subtract 1 from each element of the matrix specified in step 3.
5. Consider a random matrix key of size (digit number*digit number).
6. Multiply the output of step 4 with the output of step 5.
7. Convert all positive values of matrix to 1, negative values to -1 and zero by 0.
8. Add 1 to each element of output of step 7.
9. Convert ternary values of step 8 into decimal form. A sequence is generated.
Example
Step1:
Consider the sequence for n= 0 to 26 values.
Step2:
Convert the sequence to ternary form of a 3 digit number
i.e. 0 ------- 000
1-------- 001
2-------- 002
.
.
.
26-------- 222
Step3:
Represent above ternary form in 27x3 matrix
 0 
 1 
0 0
 0 0
 2 
 
0 0
 0 1 0 
 1 
 
0 1
 2 
 0 
0 1
 
0 2
 1 
 2 
0 2
 
0 2
 0 
 
1 0
 1 0 1 
 2 
 
1 0
 1 1 0 
=  1 
 
R 1 1
 2 
 0 
1 1
 
1 2
 1 
 
1 2
 1 2 2 
 0 
 
2 0
 2 0 1 
 2 
 
2 0
 0 
 1 
2 1
 
2 1
 2 
 0 
2 1
 2 2
 1 
 
2 2
 2 2 2 
Step 4:
Subtract 1 from each element of the above matrix and the resulting matrix R is
 − 1 − 1 − 1 
 − 1 − 1 0 

 − 1 − 1 1 
 
 − 1 0 − 1 
 − 1 0 
 
0
 − 1 1 
 − 1 
0
− 1
 
1
 − 1 0 
 1 
1
− 1
 
1
 − 1 − 1 
 
0
 0 − 1 0 
 − 1 1 
 
0
 0 0 − 1 
=  0 
 
R 0 0
 1 
 − 1 
0 0
 
0 1
 0 
 
0 1
 0 1 1 
 − 1 − 1 
 
1
 1 − 1 0 
 − 1 1 
 
1
 − 1 
 0 
1 0
 
1 0
 1 
 − 1 
1 0
 1 1
 0 
 
1 1
 1 1 1 
Step5:
Consider a random matrix
2 − 6
3 3 
5
A=  1
 4 −2 − 3 
Step6:
R= R X A
 − 1 − 7 1 
 − 7 − 4 − 2 

 − 13 − 1 − 5
 
 4 − 6 − 1
 − 2 − 31 − 4
 
 − 8 − 7
 9 − 3
0
− 5
 
 3 − 2 − 6
 − 3 − 9
 
1
 1 − 4 5 
 
 − 5 − 1 2 
 − 11 − 1
 
2
 6 − 3 3 
R =  0 0 
 
0
 − 6 − 3
 11 1 
3
− 2
 
 5 − 2
 
1
 − 1 4 − 5
 3 − 1 9 
 
 − 3 2 6 
 − 9 3 
 
5
 8 7 
 2 4 
0
 
3
 − 4 1 
 13 5 
6
 1
 7 2 
 
4
 1 7 − 1
Step7:
Convert all positive values to 1, negative values to -1 and zero to 0 of the resulting
matrix of step 6.
 − 1 − 1 1 
 − 1 − 1 − 1 

 − 1 − 1 − 1 
 
 1 − 1 − 1 
 − 1 − 1 − 1 
 
 − 1 − 1 
 − 1 
0
− 1
 
1
 − 1 − 1 
 − 1 
1
− 1
 
1
 − 1 1 
 
1
 − 1 − 1 1 
 − 1 − 1 
 
1
 1 − 1 1 
=  0 
 
R 0 0
 − 1 − 1 
 1 
1
− 1
 
1
 − 1 
 
1 1
 − 1 1 − 1 
 − 1 1 
 
1
 − 1 1 1 
 − 1 1 
 
1
 1 
 1 
1 0
 
1 1
 − 1 1 
 1 
1
 1 1
 1 
 
1 1
 1 1 − 1 
Step8: Add 1 to each element of the matrix R
 
 
0 0 2
 0 0 0 
 
 
0 0 0
 2 0 0 
 
 
0 0 0
 0 1 0 
 
 
2 0 0
 
 
2 0 0
 
0 2 0
 
 
2 0 2
 0 0 2 
 
 
0 2 0
 2 0 2 
=  
 
R 1 1 1
 
 
0 2 0
 
2 0 2
 
 
2 2 0
 0 2 0 
 
 
2 0 2
 0 2 2 
 
 
0 2 2
 
 
2 1 2
 
2 2 2
 
 
0 2 2
 2 2 2 
 
 
2 2 2
 2 2 0 
Step9:
Convert each row of the matrix R to decimal form to generate sequence.

ie 0 0 2 will form 0*3 2+0*3 1+2*3 0 = 2
The sequence formed is = 2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20

24 6 20 8 8 23 26 8 26 26 24.
Algorithm for generating Basins (Sub keys) from sequence generated.
1. Consider the sequence of values starting from 0 to n where n be an integer.

2. Read the sequence generated from algorithm 1.
3. Read the starting element of step 1 and store the first element of step 1 and the
corresponding first element of step 2 in a separate basin.
4.1 Compare the element of step 3 with the elements of step 2. If there is a
match, store the corresponding elements of step 1 in the basin specified in
step 3. Neglect already visited elements.
4.2 Repeat step 4.1 with the remaining elements of the basin of step 3 and
store them in the same basin. This will form one basin.
5. Go to next element of step 1 which is not visited earlier.
For example
1. n [27]= 0 1 2 3 4 5……………………………………………...26
2. r [27] = 2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20 24 6 20 8
8 23 26 8 26 26 24.
3. Read n[0]=0. Store the values of n[0],r[0] in a basin. ie b(0)=(0,2).
4.1. n[0] ie. ‘0’ is compared with r[27] values. There is a match at r[1],r[2] & r[4].
Neglect already visited elements. Thus b(0)=[0,1,2,4).
4.2 Step 4.1 is repeated with other elements of basin ie. 1, 2 & 4 values. For
elements 1& 4, there is no match of values in r[27]. For element 2, there is a
match at r[10]. Thus the basin b(0) is updated to (0, 1, 2, 4&10).
5. The procedure is repeated for the next element of the sequence of step 1 which is not
visited earlier. The other basins formed are
b(1)=(3,5,6,7,8,9,11,12,14,15,17,18,19,20,21,23)
b(2)=(13)
b(3)=(16,22,24,25,26)
Mapping the sequence or the basins on the plain text to generate three
models of cipher text.
3.2.1 Model 1:
A new substitution technique
The algorithm [22, 34] considers a matrix key and executes a sequence of steps
which generates a sequence. Each block of plain text is replaced by summation of
alphanumerical value of plain text and the sequence generated to form cipher text. Thus
the cipher text obtained becomes very difficult to break with out knowing the key.
3.2.2 Model 2:
A New Variable Length Key Block Cipher Technique for Network Security in Data
Transmission[15].
The algorithm considers a random matrix key which on execution of a sequence
of steps generates a sequence. Based on the equality of values this sequence is being
divided into basins. Each basin represents one block of data. Depending on starting input
plain text character, corresponding basin is considered as a key. Each block of plain text
is replaced by summation of alphanumerical value of plain text and the sequence
generated to form cipher text. The procedure is repeated for certain plain text depending
on chosen value. Thus the cipher text obtained becomes very difficult to be broken with
out knowing the key. In a variable length encryption scheme, the key used is of variable
length. The scheme requires one time execution of the key schedule to generate all sub
keys prior to encryption. This provides for slight resource burden on the key agility.
3.2.3 Model 3:
A New probabilistic Stream encryption scheme
leaked out. Because a crypto analyst can always encrypt random messages with a public
key, he can get some information. Not a whole of information is to be gained here, but
there are potential problems which allow a crypto analyst to encrypt random messages
with public key. Some information is leaked out every time to the crypto analyst, he
encrypts a message. In probabilistic multiple cipher texts are generated for one plain text,
a cryptanalyst can not generate any information by chosen plain text and chosen cipher
text attacks.
The algorithm [23] considers a matrix key and executes a sequence of steps which
generates a sequence. Based on the equality of values this sequence is being divided into
basins. The basins with minimum values will be neglected. Remaining basins represent
the characters of the alphabet. Each plain text character is replaced by a set of basins.
Each basin value is replaced by random value from that basin to generate cipher text.
Thus multiple cipher texts will be generated for one plain text and any one cipher text can
be used for transmission of data. Thus the cipher text obtained becomes computationally
infeasible to be broken with out knowing the key.
Thus the algorithm generates a subkey which will be used for encryption and
decryption processes used to develop three models. Model 1 and Model 2 are block
ciphers and Model 3 is stream cipher. Encryption & Decryption Processes are inverse to
one another which make the same algorithm to be used for encryption and decryption
process with a difference of addition in Encryption with subtraction in Decryption
Process.
A NEW SUBSTITUTION BLOCK CIPHER
Model 1
3.2.1 A new Substitution cipher for data security.
The algorithm that is going to be discussed in this work will generate a Sequence.
The algorithm
m considers a matrix key and executes a sequence of steps which generates the
sequence. Each block of plain text is replaced by summation of alphanumerical value of
the plain text and the sequence generated to form cipher text. Thus the cipher text
obtained becomes computationally infeasible to break with out knowing the key
3.2.1.1 Model 1: A new Substitution cipher for data security.
N = 0-26. Consider key Apply a sign A sequence is Sequence is

Convert n to Multiply it function on the generated mapped to
with ternary output. plain text to
ternary vector generate
form cipher text.
3.2.1.2 The new algorithm has the following features…

1. A set of poly alphabetic substitution rule is used.
2. A new block cipher is developed.
3. A Random matrix is being used as key.
4. Generated sequence being used as sub key.
5. Coding method.
The steps that are involved in the proposed algorithm.

1. The decimal values and letters of the plain text are given numerical values
starting from 0
2. A random matrix is used as a key. Let it be A.
3. A “ternary vector” for 33 values i.e from 0 to 26 is generated
4. Let this be “B”.
5. 1 is subtracted from all the values of ternary vector.
6. The modified ternary vector is multiplied with the matrix key.
7. A sign function is applied on the product of ternary vector & matrix key.
8. 1 is added to all values of step 7.
9. A sequence is generated which is used as sub key
10. The sub key is added to the individual numerical values of the message to
generate cipher text.
It can be seen that to extract the original information from the coded text is highly
impossible for the third person who is not aware of encryption keys and the method of
coding.
Even if the algorithm is known it is very difficult to break the code and generate
key, given the strength of the algorithm .Thus given a short response time through
internet communication, the algorithm is supposed to be safe.
3.2.1.3 Some of the advantages of this algorithm can be summarized as

follows.
1. Even if the algorithm is known, it is computationally infeasible to generate
matrix key.
2. Versatile to users: Different users of internet can use different modified
versions of the new algorithm. Since in this algorithm, the sign function is used, it
is supposed to be strong enough.
3. As per the matrix the same character is substituted by different alpha numerical
values which provide more security for the message.
3.2.1.4 Example:
Sequence generated from the product of matrix key considered and ternary vector (ref.
pp 44 ) = 2 0 0 18 0 3 18 18 6 20 10 ……………………………. 26 24.
Table No. 3.2.1.1: Encryption & Decryption process by Model 1:

Encryption.
Plain Text A v n K R I
Alpha 10 31 23 20 27 18
Numeric
equivalent
Key 2 0 0 18 0 3
Add 12 31 23 38 27 21
Mod 36 12 31 23 02 27 21
Cipher C v n 02 R L
Text
Decryption
Cipher C v n 02 R L
Text
Alpha 12 31 23 02 27 21
Numeric
equivalent
Add 36 12 31 23 38 27 21
Key 2 0 0 18 0 3
Subtract 10 31 23 20 27 18
Plain Text A v n K R I
3.2.1.5 Computing Power
Total number of computations considered in the given model for converting plain text to
cipher text.
Computation one: Converting n=0:26 to ternary vector. Let it be r.
Computation two: Calculating r-1 and storing it in r.
Computation three: Multiplying r with the key considered.
Computation four: Applying sign function on the product. Store it in r.
Computation five: Calculating r+1
Computation six: Converting output ternary vector to integer form. Let this be s,
the sequence generated.
Computation seven: Converting plain text to alphanumerical value.
Computation eight: Adding alphanumerical value of plain text to sequence
generated.
Computation nine: Applying mod function on the output.
Computation ten: Converting the output to characters of the alphabet to get
cipher text.
Thus the total number of computations in the first proposed model is 10.
Computation overhead (Computing Power) for a 9 character key
1st computation: 27 calculations, 2nd computation: 27 calculations.

Key considered: 9 character key, 3rd computation: 27*9 calculations, 4th computation:
27 calculations, 5th computation: 27 calculations, 6th computation: 27 calculations, 7th
computation: 27 calculations. Considering a 27 character plain text, 8th computation: 27
calculations, 9th computation: 27 calculations, 10th computation: 27 calculations.
Thus the total computational overhead by the first model is 486 calculations
3.2.1.6 Complexity of the Model 1

Complexity by Construction
1st Computation: Converting n=0:26 to ternary vector. Let it be r. The
complexity is in multiples of n.
2nd Computation: Calculating r-1. The complexity is in multiples of n
3rd Computation: Multiplying r with the key considered. The complexity is in
multiples of n
4th Computation: Applying sign function on the product. Store it in r. The
complexity is in multiples of n
5th Computation: Calculating r+1. The complexity is in multiples of n
6th Computation: Converting output ternary vector to integer form. Let this be s,
the sequence generated. The complexity is in multiples of n
7th Computation: Converting plain text to alphanumerical value. The
8th Computation: Adding alphanumerical value of plain text to sequence
generated. The complexity is in multiples of n
9th Computation: Applying mod function on the output. The complexity is in
multiples of n
10th Computation: Converting the output to characters of the alphabet to get
cipher text. The complexity is in multiples of n.
Thus we can say that the complexity of model1 is O (n).
Complexity of the algorithm by its strength
Complexities can also be expressed as orders of magnitude. If the length of the
key is k, then the processing complexity is given by 2k. It means that 2 k operations are
required to break the algorithm. In the given algorithm a matrix key is used. This matrix
key is multiplied with ternary vector. On the generated values a sign function is used to
convert all positive values to 1, negative values to –1 and zero to 0. This provides the
necessary strength to the algorithm. Thus known the algorithm, known the cipher text it is
quite difficult to generate the matrix key. Thus in the present algorithm, there is no means
by which the key can be retrieved, other than trying all the combinations of key, the
complexity of the algorithm is said to be exponential in nature.
.
Table No.3.2.1.2: Representation of No. of computations, Computing power,
Complexity by construction & Complexity by strength
S No. Algorithm Number of Computing Complexity Complexity

computations Power by by Strength
Construction
1 Converting 1 27 N ---
n=0:26 to
ternary vector.
Let it be r.
2 calculating r-1 1 27 N ---
3 multiplying r 1 27*9 N ---
with the key
considered
4 Applying sign 1 27 N ---
function on the
product
5 calculating r+1 1 27 N ---
6 converting 1 27 N ---
output ternary
vector to
integer form
7 converting 1 27 N ---
plain text to
alphanumerical
value
8 Adding 1 27 N ---
alphanumerical
value of plain
text to
sequence
generated
9 Applying mod 1 27 n ---
function on the
output
10 converting the 1 27 n Exponential

output to
characters of Total=10 Total=486. O(n)
the alphabet to
get cipher text
3.2.1.7 Avalanche effect:

In this model a sequence is generated and this sequence is substituted for the plain
text to generate cipher text. Depending on the key, the sequence will be generated. We
will identify the variations in the sequence generated, by slight variations in the key .Thus
we can identify the variations in the cipher text by slight variations in the key considered.
We will also identify the variations in the cipher text by slight variations in the plain text.
For example, considering different cases for slight variations in the key,
Case 1:
2 5 −6
A=key 3 1 3
4 −2 −3
Sequence generated from the proposed model
n=0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
r=2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20 24 6 20
n= 19 20 21 22 23 24 25 26
r= 8 8 23 26 8 26 26 24.
Case 2: By increasing the values of key by 1 at each row

3 5 −6
A = key = 4 1 3
5 −2 −3
Sequence generated
r= 1 0 0 18 0 0 18 18 3 20 2 6 20 13 6 20 24 6 23 8 8 26 26 8 26 26 25.
Case 3:
By decreasing the key values by 1 at each row.
1 5 −6
A = key = 2 1 3
2 −2 −3
Sequence generated by the model.
r= 11,1,3,20,0,6,18,18,6,20,2,6,20,13,6,20,24,6,20,8,8,20,26,6,23,25,15.
Table No. 3.2.1.3: Identifying the variations in cipher text by slight variations in the
key.
(Model 1)
Case Considered Key Considered Plain text considered Cipher text formed
Case 1 [2 5 -6; 3 1 3; 4 - a b c d e cb cv e
2 -3]
Case 2 [3 5 -6; 4 1 3; 5 - a b c d e bbc v e
2 -3]
Case 3 [1 5 -6; 2 1 3; 2 - a b c d e i c f x e
2 -3]
` Thus we can see that, by changing the key slightly, there are a lot of variations in
the cipher text which provides maximum avalanche effect to the algorithm. This provides
for maximum strength and security to the algorithm. But since the model is a simple
substitution algorithm, the plain text variations to cipher text variations is negligible.
3.2.1.8 Security Analysis:
The model uses a sign function on the product of ternary vector and a matrix key
to generate the sequence. The sign function converts all positive values to 1, negative
values to -1, and zero with 0.This sequence is substituted for plain text to generate cipher
text. Thus it is impossible to generate the matrix key from the known plain text and
cipher texts. Thus this model is free from differential crypto analysis.
But this model uses a simple substitution technique to generate cipher text; it is
some what susceptible to linear crypto analysis. The Key can not be gained and not a
whole of information can be gained, but part of information may be gained in this model.
This algorithm is completely free from cipher text only, type of attack. By the other
attacks, the key may not be retrieved but a part of plain text may be retrieved.
3.2.1.9 Conclusion:
In this work a ternary system with a 3 digit number is used. So the sub key
generated is a 3 3 ie a 27 digit number. By considering a ternary vector with a four digit
number or five digit number, the length of the sub key can be increased by 34, 35 which
increases the length of sub key generated. Similarly by considering n –ary vector the
length of the subkey generated can still be increased. Thus by increasing the length of
subkey, security of cipher system can be increased still further.
A NEW VARIABLE LENGTH KEY BLOCK CIPHER
Model 2
3.2.2 A New Variable Length Key Block Cipher Technique for Network
Security in Data Transmission
The algorithm that is going to be discussed in this work is going to consider a
random matrix key which on execution of sequence of steps generates a sequence. Based
on the equality of values this sequence is being divided into basins. Each basin represents
one block of data. Depending on starting input plain text character, corresponding basin
is considered as a key. Each block of plain text is converted to alphanumerical values
which are mapped with the sub key to generate cipher text. The procedure is repeated for
certain plain text depending on chosen value. Thus the cipher text obtained becomes very
difficult to be broken with out knowing the key.
Key Words: Cryptography, Variable length key, Encryption Algorithm, Example, Add
function.
3.2.2.1 Model 2: A new variable length key for data encryption.
N=0-26. Consider key Apply a sign A sequence Similar Basins are

Convert n to Multiply it function on is generated values of mapped to
ternary form with ternary the output. sequence are plain text to
vector stored in form cipher
basins. text .
3.2.2.2 A New Algorithm: Model 2
The new algorithm has the following features…

1. A new block cipher is developed using a variable length key.
2. A Random Matrix is used as a key.
3. Based on equality of values, the sequence is divided into basins.
4. These basins (sub keys) are used to the characters in the plain text by a
particular chosen rule.
5. Coding method.
Thus this algorithm is a combination of
1. A ternary vector of 27 values to few basin values.
2. Each basin containing different number of values.
3. Corresponding basin is considered as key depending on first character of plain
text.

1. The decimal values and letters of the alphabet are given numerical values
starting from 0.
3. A “ternary vector” for 33 values i.e from 0 to 26 is generated.
4. Let this be “B”.
5. 1 is subtracted from all the values of step 4.
6. The generated ternary vector is multiplied with matrix key.
7. A sign function is applied on the product of ternary vector & the matrix key.
8. 1 is added to all values of step 7
9. A sequence is generated.
10. Basins will be developed from this sequence which contains equal values.
11. Basins with minimum values (say 1 value) are eliminated.
12. The first value of plain text is considered.
13. The remainder of the first value of plain text with the number of basins is
calculated.
14. Depending on the value of the remainder, the corresponding basin is
considered as a key.
15. The key is added to the plain text to generate cipher text.
16. The procedure is repeated for successive plain texts depending on chosen
value.
3.2.2.3 Advantages
1. It is almost impossible to extract the original information.
2. Even if the algorithm is known, it is difficult to extract the matrix key.
3. Versatile to users. Different users of internet can use different modified
versions of the new algorithm.
4. As per basin values, the same character is substituted by different alpha
numerical value which provides more security for the message.
3.2.2.4 Example:
The sequence generated from the product of ternary vector & Matrix key
considered (Ref. pp 44 )
2 0 0 18 0 3 18 18 6 20 10 ……………………………. 26 24.
The basins that can be formed using this sequence are
b(0)=(0,1,2,4,10)
b(1)=(3,5,6,7,8,9,11,12,14,15,17,18,19,20,23)
b(2)=(16,21,22,24,25,26)
Plain text Considered: avnkri

Considering the first character of the plain text, a = 10, 10 mod 3=1. So the basin
considered is b(1)
Table No.3.2.2.1 Encryption & Decryption Process by Model 2:
1. Encryption
Plain text A V N K R I
Alpha 10 31 23 20 27 18
Numeric
Equivalent
Key(Basin 3 5 6 7 8 9
Considered)
Add 13 36 29 27 35 27
Mod 36 13 0 29 27 35 27
Cipher Text D 0 T R Z R
2. Decryption
Cipher Text D 0 T R Z R
Alpha 13 0 29 27 35 27
Numeric
Equivalent
Key(Basin 3 5 6 7 8 9
Considered)
Subtract 10 -5 35 20 27 18
Add 36, if 10 31 35 20 27 18
negative
Plain text A V n K R I
After some part of plain text, the procedure is repeated.
Total number of computations considered in the given model for converting
plain text to cipher text.
Computation two: Calculating r-1.
Computation seven: Generating different basins by placing equality of values of
the sequence in one basin.
Computation eight: Converting plain text to alphanumerical value.
Computation nine: Considering the first character of plain text and applying a
mod function of the order of number of basins formed.
Computation ten: Depending on the output of the mod function, the
corresponding basin is used as a key.
Computation eleven: Adding the key to the alphanumerical value of the plain
text.
Computation twelve: Applying mod function on the output.
Computation thirteen: Converting the output to characters of the alphabet to get
cipher text.
Computation fourteen: Repeating the procedure for chosen part of plain text.
Thus the total number of computations in the proposed model is 14.
Computation overhead of the Model with a 9 character key

1st computation: 27 calculations, 2nd computation: 27 calculations
Key considered: 9 character key, 3rd computation: 27*9 calculations, 4th
computation: 27 calculations, 5th computation: 27 calculations, 6th computation:
27 calculations, 7th computation: 27*27 calculations +27*27 calculations , 8th
computation: 27 calculations, 9th computation: 03 calculations depending on
chosen rule, 10th computation: 03 calculations, 11th computation: 27 calculations,
12th computation: 27 calculations, 13th computation: 27 calculations, 14th
computation: 03 calculations depending on the chosen rule.
Thus the total computational overhead by the second model is 1953 calculations
3.2.2.6 Complexity of the Model

Complexity of the Model by Construction
3rd computation: Multiplying r with the key considered. The complexity is in
multiples of n
7th Computation: Generating basins from sequence which involves 27*27
calculations +27*27 calculations. The complexity is in multiples of n*n.
8th Computation: Converting plain text to alpha numerical value ie 27
calculations. The complexity is in multiple of n
9th Computation: Considering the first character of plain text and applying a
mod function of the order of number of basins formed.03 calculations depending
on chosen rule. The complexity depends on chosen rule.
10th Computation: Depending on the output of the mod function, the
corresponding basin is used as a key ie 03 calculations. The complexity depends
on chosen rule.
11th Computation: Adding the key to the alphanumerical value of the plain text
ie 27 calculations. The complexity is in multiples of n.
12th Computation: Applying mod function on the output ie 27 calculations. The
13th Computation: Converting the output to characters of the alphabet to get
cipher text.
ie 27 calculations. The complexity is in multiples of n.
14th Computation: Repeating the procedure for chosen part of plain text ie 03
calculations depending on the chosen rule. The complexity depends on chosen
rule.
Thus the total computational complexity of the model by its construction is of the order
of O (n 2)

convert all positive values to 1, negative values to –1 and zero to 0. This provides the
necessary strength to the algorithm. Thus known the algorithm, known the cipher text it is
quite difficult to generate the matrix key. Thus in the present algorithm, there is no means
by which the key can be retrieved, other than trying all the combinations of key, the
complexity of the algorithm is said to be exponential in nature. Thus this algorithm is
supposed to be safe in real time environment.
Table No.3.2.2.2 Representation of No. of computations, Computing power,
Complexity by construction & Complexity by strength
Algorithm Number of Computing Complexity Complexity
S No. computations Power by by Strength
Construction
n=0:26 to
ternary vector.
Let it be r.
2 calculating r-1 1 27 N ---
3 multiplying r 1 27*9 n ---
with the key
considered
4 applying sign 1 27 n ---
function on the
product
5 calculating r+1 1 27 n ---
6 Converting 1 27 n ---
output ternary
vector to
integer form
7 Dividing the 1 27*27 n2 ---
sequence to
Basins
8 Converting 1 27 n ---
plain text to
alphanumerical
value
9 Identifying 2 27 n ---
basin as sub
key
10 Adding 1 27 n ---
alphanumerical
value of plain
text to sub key
11 applying mod 1 27 n ---
function on the
output
12 converting the 2 27 n Exponential

output to
characters of Total=14 Total=1953 O(n 2)
the alphabet to
get cipher text
In this model a sequence is generated and this sequence is used to generate basins of
variable length. These basins will be used as keys for the plain text to generate the cipher
text. We will identify the variations in the basins generated, by slight variations in the key
.Thus we can identify the variations in the cipher text by slight variations in the key
considered. We will also identify the variations in the cipher text by slight variations in
the plain text.
For example, considering different cases for slight variations in the key,
Case 1:
2 5 −6
A=key 3 1 3
4 −2 −3
n=0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
r=2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20 24 6 20
n= 19 20 21 22 23 24 25 26
r= 8 8 23 26 8 26 26 24.
the basins that can be formed using this sequences which are used as sub keys are
b(0)=(0,1,2,4,10)
b(1)=(3,5,6,7,8,9,11,12,14,15,17,18,19,20,21,23)
b(2)=(16,22,24,25,26)
3 5 −6
A = key = 4 1 3
5 −2 −3
Sequence generated
r= 1 0 0 18 0 0 18 18 3 20 2 6 20 13 6 20 24 6 23 8 8 26 26 8 26 26 25.
Thus the basins formed which are used as sub keys are
b(0)= (0,1,2,4,5,10)
b(1)= (3,6,7,8,9,11,12,14,15,17,18,19,20,23)
b(2)= (16,21,22,24,25,26);
Case 3:
By decreasing the key values by 1 at each row
1 5 −6
A = key = 2 1 3
2 −2 −3

r= 11,1,3,20,0,6,18,18,6,20,2,6,20,13,6,20,24,6,20,8,8,20,26,6,23,25,15.
Basins formed which are used as sub keys are
b(0) = (0,11,4)
b(1)= (2,3,10)
b(2)= (5,6,7,8,9,11,12,14,15,17,18,19,20,21,22,23,24,26)
Note: Considering the basins as keys, by neglecting the basins which contain minimum
number of values, say 1.
Table No. 3.2.2.3 Identifying the variations in cipher text by slight variations in the
key (Model 2)
Case considered SubKey considered Plain text Cipher text

as per the first
character of plain
text
Case 1 B(1)=3 5 6 7 8 abcde dgikm
Case 2 B(1)=3 6 7 8 9 abcde dhjl o
Case 3 B(1)=2 3 10 2 3 abcde cemfh
Thus we can see that, by changing the key slightly, there is a lot of variations in
the basins generated. Since these basins will form the sub keys for the given model, they
provide maximum avalanche effect to the algorithm. Thus this model provides good
variations in cipher text for slight variations in the key. This provides for maximum
strength and security to the algorithm.
Table No.3.2.2.4 Identifying the variations in cipher text by slight variations in the
plain text (Model 2)
Case considered SubKey considered Plain text Cipher text
as per the first
character of plain
text
Case 1 36789 abcde dhjln
Case 1 16 22 24 25 26 bbcde hortu
Case 1 0 1 2 4 10 cbcde aceho
We will see a lot of variations in the cipher text generated for slight variations in the plain
text. Thus it provides a maximum avalanche effect to the algorithm which provides for
more strength and security.

cipher texts. Thus this model is free from differential crypto analysis. Since variable
length keys are used it is also free from linear crypto analysis.
This algorithm is completely free from cipher text only and known plain text attacks
but may not be completely free from chosen plain text and cipher text attacks where a
part of information may be leaked out by cryptanalysis.
3.2.2.9 Conclusion:
In the given work, a ternary vector with a 3 digit number is used. By using a n-ary
vector the length of the vector can be increased. By increasing the length of the vector,
the number of basins may not be increased. But the number of values in each basin will
be increased which provides more strength to the developed model. It is also observed
that by slight variations in the key values and plain text, the number of basins formed and
number of values of each basin are varying in nature, which provides a better avalanche
effect. This provides more security and strength to the algorithm
A NEW PROBABILISTIC STREAM CIPHER
Model 3
3.2.3 A New probabilistic encryption scheme:
With probabilistic encryption algorithms, a crypto analyst can no longer encrypt
random plain texts looking for correct cipher text. Since multiple cipher texts will be
developed for one plain text, even if he decrypts the message to plain text, he does not
know how far he had guessed the message correctly. To illustrate, assume a crypto
analyst has a certain cipher text ci. Even if he guesses message correctly, when he
encrypts message the result will be completely different cj. He cannot compare ci and cj
and so cannot know that he has guessed the message correctly. Under this scheme,
different cipher texts will be formed for one plain text. Also the cipher text will always be
larger than plain text.
The new encryption algorithm is based on the concept of Polyalphabet cipher

which is an improvement over monoalphabetic technique. In this algorithm one character
of plain text is replaced by set of values from different basins. Thus a polyalphabet is
maintained.
The algorithm that is going to be discussed in this work is going to generate a

Sequence. The algorithm considers a matrix key and executes a sequence of steps which
generates this sequence. Based on the equality of values this sequence is being divided
into basins. The basins with minimum values will be eliminated. From the remaining
basins each basin represents one character. Each Character in the plain text is replaced
by a corresponding basin value. Thus the cipher text obtained becomes impossible to be
broken with out knowing the key.
3.2.3.1 Model 4: A New probabilistic encryption scheme:

N=0-26.
Convert n to
ternary form
Consider key
Multiply it
with ternary
vector
Apply a sign
function on
the output.
A sequence is
generated.
Similar
values of
sequence are
stored in
basins.
Each
character of
plain text are
represented
by set of
basins.
Values from
basins are
randomly
represented to
form cipher
text.
3.2.3.2 The new algorithm has the following features
1. A new probabilistic stream cipher is developed.
2. A matrix is used as a key.
3. A sequence is generated by the product of matrix key & the ternary vector.
4. This sequence is used to generate basins based on equal values.
5. Each plain text character is replaced by random values from basins to generate
cipher text.
6. Coding method.
Thus this algorithm is a combination of

1. A ternary vector of 27 values to few basin values.
2. Each basin containing different number of values.
3. Each character of the alphabet being represented by a set of basins.

1. The letters of the alphabet are given numerical values starting from 1.
3. A “ternary vector” for 33 values i.e from 0 to 26 is generated.
4. 1 is subtracted fro all values of step 3.
5. The generated ternary vector is multiplied by the matrix key.
6. A sign function is applied on the product of ternary vector & the matrix key.
7. 1 is added to all values of step 6.
8. A sequence is generated.
9. Basins will be developed from this sequence which contains equal values.
10 Each plain text character is converted to set of basins based on chosen base
value.
11 Each basin is replaced by values from basins in random form to generate the
cipher text.
3.2.3.3 Advantages:
1. It is almost impossible to extract the original information.
2. Even if the algorithm is known, it is difficult to extract the key.
3. Versatile to users. Different users of internet can use different modified
versions of the new algorithm. Since in this algorithm all positive values are
consider as +1 and all negative values are consider as -1 and zero as 0,it is
impossible to generate the matrix key even if plain text and cipher text are known
.
4. As per the matrix the same character is substituted by different alpha numerical
value which provides more security for the message.
5. By suitably combining basins, the number of characters of the alphabet can be
increased.
3.2.3.4 Example:
For n= 0 : 26 ;
Sequence generated r (ref. pp 43)=2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20 24 6
20 8 8 23 26 8 26 26 24.
The basins that can be formed using this sequences are
b(0)=(0,1,2,4,10)
b(1)=(3,5,6,7,8,9,11,12,14,15,17,18,19,20,21,23)
b(2)=(16,22,24,25,26)
Now considering a base value of 3 i.e. representing a character with random values from
3 basins, the total characters that can be represented in this alphabet is 3 3 =27.
For example if K=11, it can be represented as 1 0 2 = b1 b0 b2.
Table No. 3.2.3.1 Encryption & Decryption Process by Model 3.
Encryption
Plain text A S K
Alpha numeric 01 19 11
equivalent
Equivalent basins b(0)b(0)b(1) b(2) b(0) b(1) b(1) b(0) b(2).
Random values from 1 2 8 24 4 9 12 2 26

basins
Cipher text C i a b h x d i l b z
Random values from 2 2 16 26 1 14 16 4 24

basins
Cipher text Cj b b p z a n p d x
Decryption
Cipher text C i a b h x d i l b z
Alpha numeric 1 2 8 24 4 9 12 2 26
equivalent
Equivalent basins b(0)b(0)b(1) b(2) b(0) b(1) b(1) b(0) b(2).
Alpha numeric 01 19 11
equivalent
Plain text A S K
Total number of computations considered in the given model for converting plain text to
cipher text.
Computation two: Calculating r-1.
Computation seven: Generating different basins by placing equal values of the
sequence in one basin.
Computation eight: Converting plain text to alphanumerical value.
Computation nine: Representing each character by a set of basins based on
chosen base value.
Computation ten: Representing the basins by their corresponding values in a
random manner.
Computation eleven: Converting the output to characters of the alphabet to get
cipher text.
Thus the total number of computations in the first proposed model is 11.
Computation overhead for a 27 character key
1st computation: 27 calculations, 2nd computation: 27 calculations
Key considered: 9 character key, 3rd computation: 27*9 calculations, 4th computation:
27 calculations, 5th computation: 27 calculations, 6th computation: 27 calculations, 7th
computation: 27*27 calculations +27*27 calculations, 8th computation: 27 calculations,
9th computation: 81 calculations, 10th computation: 81 calculations, 11th computation:
81calculations + Number of computations for random choice of values from basins (k).
Thus the total computational overhead by the third model is 2106+k calculations
3.2.3.6 Complexity of the Model
Complexity of the algorithm by construction

3rd Computation: Multiplying r with the key considered. The complexity is in
multiples of n
7th Computation: Generating Basins ie 27*27 calculations +27*27 calculations.
The complexity is in multiples of n*n.
8th Computation: Converting plain text to alpha numerical value ie 27
calculations. The complexity is in multiples of n
9th Computation: Converting plain text by set of basins ie 27*3 calculations.
The complexity is in multiples of n
10th Computation: Representing basins by values from basins randomly ie 27*3
calculations. The complexity is in multiples of n
11th computation: Converting values from basins to cipher text ie 27*3
calculations. The complexity is in multiples of n.
Thus the total computational complexity of the model by its construction is of the order
of O (n 2)
convert all positive values to 1, negative values to –1 and zero to 0. A sequence is
generated. This sequence is divided to basins depending on similar values. The characters
of plain text are replaced by set of basins based on chosen base value. The basins are
replaced by random values of the corresponding basins. This provides the necessary
strength to the algorithm. Thus known the algorithm, known the cipher text it is quite
difficult to generate the matrix key. Thus in the present algorithm, there is no means by
which the key can be retrieved, other than trying all the combinations of key, the
complexity of the algorithm is said to be exponential in nature. Thus this algorithm is
supposed to be safe in real time environment.
Table No.3.2.3.2 Representing No. of computations, computing power, complexity
by construction & strength
S No. Algorithm Number of Computing Complexity Complexity
computations Power by by Strength
Construction
n=0:26 to
ternary vector.
Let it be r.
2 Calculating r-1 1 27 N ---
3 Multiplying r 1 27*9 N ---
with the key
considered
4 Applying sign 1 27 N ---
function on the
product
5 Calculating r+1 1 27 N ---
output ternary
vector to
integer form
7 Dividing the 1 27*27+27*27 n 2 ---
sequence to
Basins
plain text to
alphanumerical
value
9 Converting 1 27*3 N ---
Plain text value
by set of basins
10 Representing 1 27*3+k N ---
values fro
basins in
random order
11 Generating 1 27*3 N Exponential
cipher text
Total O(n 2)
=2106+k
In this model a sequence is generated and this sequence is used to generate basins
of variable length. These basins will form the characters of the alphabet. A combination
of basins will be formed to increase the characters of the alphabet. This combination of
basins will be mapped for the plain text to generate the cipher text. We will identify the
variations in the basins generated, by slight variations in the key .Thus we can identify
the variations in the cipher text by slight variations in the key considered. We will also
identify the variations in the cipher text by slight variations in the plain text. For example,
For example, considering the key A of size(3*3)

Case 1:
2 5 −6
A=key 3 1 3
4 −2 −3
n=0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
r=2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20 24 6 20
n= 19 20 21 22 23 24 25 26
r= 8 8 23 26 8 26 26 24.
thus the basins that can be formed using this sequences are
b(0)=(0,1,2,4,10)
b(1)=(3,5,6,7,8,9,11,12,14,15,17,18,19,20,21,23)
b(2)=(13)
b(3)=(16,22,24,25,26)
3 5 −6
A = key = 4 1 3
5 −2 −3
Sequence generated
r= 1 0 0 18 0 0 18 18 3 20 2 6 20 13 6 20 24 6 23 8 8 26 26 8 26 26 25.
Thus the basins formed are
b(0)= (0,1,2,4,5,10)
b(1)= (3,6,7,8,9,11,12,14,15,17,18,19,20,23)
b(2)=(13)
b(3)= (16,21,22,24,25,26);
Case 3:
1 5 −6
A = key = 2 1 3
2 −2 −3
r= 11,1,3,20,0,6,18,18,6,20,2,6,20,13,6,20,24,6,20,8,8,20,26,6,23,25,15.
Basins formed are

b(0) = (0,11,4)
b(1) = (1);
b(2)= (2,3,10);
b(3)= (5,6,7,8,9,12,14,15,17,18,19,20,21,22,23,24,26)
b(4)=(13)
b(5)=(25).
key
Considering Number of sub keys Number of Ratio of number of
different cases formed characters of the characters of cipher
alphabet which the text to plain text
cipher text supports
Case 1 4 64 3
Case 2 4 64 3
Case 3 6 36 2
Considering Plain text Cipher text Data overhead

different cases
Case 1 a b c bdr0axjfb 3 times
Case 2 a b c Ej00bidpa 3 times
Case 3 a b c 0akbde 2 times
plain text
different cases
Case 1 a b c 0biabpacj 3 times
Case 1 b b c b0v0jpj0d 3 times
Case 1 c b c 0s0bdxacj 3 times
Table No. 3.2.3.5 Identifying multiple copies of cipher text for one plain text
different cases
Case 1 a b c 0ahbap0kj 3 times
Case 1 a b c 0jwdavjrd 3 times
Case 1 a b c Jbnbdyatd 3 times
Thus a lot of variation in the cipher text is observed by slight variations in the key
and the plain text. This provides for maximum avalanche effect on the algorithm which
provides maximum strength and security to the algorithm. And also it is observed that by
slight variations in the key, there is a substantial variation in the number of basins formed.
This will reduce the data overhead as less combination of basins will be sufficient to
support the characters of the alphabet.
cipher texts. Thus this model is free from differential crypto analysis and since basins
with different and variable values are mapped to the characters of plain text to form the
cipher text, it is also free from linear crypto analysis.
This algorithm is completely free from cipher text, known plain text, chosen plain
text, chosen cipher text attacks. This algorithm is also free from public key attacks.
3.2.3.9 Conclusion:
In this work a ternary system with a 3 digit number is used. So the sequence
generated is a 27 digit number. By using a n-ary vector, the length of the vector can be
further increased. But this does not guarantee in the increase in number of basins formed.
The number of values of generated sequence will increase in each basin. Thus the new
model provides a new probabilistic substitution mechanism where each character of plain
text is replaced by two or three characters of cipher text depending on the chosen key.
And also the model develops multiple cipher texts for one plain text. The algorithm
provides almost equal security at low computational overhead. And also the given
algorithm is free from differential and linear crypto analysis, which makes it suitable in
data encryption.
The limitation with this algorithm is more data has to be transmitted than the
actual data which demands for more band width requirements.
3.3 A New Mathematical Model with a key & a Time stamp for
encryption mechanism.
In this work a model is going to be used which develops data distributed over a
identified value which is used as nonce (IV). By properly considering an empirical
value, data is derived from the developed model. This empirical value is considered
as a key. The process is repeated for different timings which are used as time stamps
in the encryption mechanism [11,33]. Thus this model generates a distributed
sequence which is used as sub key. The model involves a identified value which is
used as nonce (IV), a considered empirical value which is used as key and different
timing as time stamps which are very important parameters in symmetric data
encryption schemes which supports not only security but also timeliness of
encryption mechanism and also acknowledgement between the participating parties.
3.3.1 Mathematical modeling of the problem
The approach to time series analysis was the establishment of a mathematical model
describing the observed system. Depending on the appropriation of the problem a
linear or nonlinear model will be developed. This model can be useful to generate
data at different times to map it with plain text to generate cipher text.
3.3.2 Linear data flow Problem.

The inintialization vector (IV) considered in the problem is n.
When t=0, T (I) =Y (I) =300. where I=1,2,……..M.
Dividing the problem area into M number of points, and for simplicity by assuming
data of the first and Mth grid points are considered to be known and constant.
The conservation equation with respect to data , time & space is represented as
α ∂ 2T/ ∂x 2 =∂T/ ∂t where α is a constant, T is the data value, x is
the space value & t is time value [38].
Integrating on both sides
For the grid points 2, M-1, the coefficients can be represented by considering the
conservation equation,
(α/∂.x) *(T I+1 n+1- T I n+1 ) – (α/ ∂x)* (TI n+1- T I-1 n+1 ) = (∂ x /∂ t)* ( T I n+1 –T I n ).
(α*∂ t /(∂.x) 2 )*(T I+1 n+1- T I n+1 ) – (α*∂ t /(∂.x) 2 )*(TI n+1- T I-1 n+1 ) = ( T I n+1 –T I n
).
(-2*α*∂ t /(∂.x) 2 )*(T I n+1 ) + (α*∂ t /(∂.x) 2 )*( T I-1 n+1 ) + (α*∂ t /(∂.x) 2 )*(T I+1 n+1)
= ( T I n+1 –T I n ).
n+1
where T I represents data value for the considered grid point for the current delt,
n+1 n+1
T I-1 &T I+1 represents data values for the preceding and succeeding grid
points for the current delt, T I n represents data item for past delt, α be a constant.
Considering α which is a key for the given model, the coefficients are obtained for
each state (grid point) in terms of A(I) refers to data value of the corresponding grid
point, C(I) and B(I) refers to data values of preceding and succeeding grid points for
the current delt, D(I) refers to data value of the considered grid point in the preceding
delt.
A(I)= 1 + 2 α delt/(delx)**2.
B(I)= -α delt/(delx)**2.
C(I)= - α delt/(delx)**2.
D(I)= T I n
3.3.3 Procedure for generating data from coefficients by tridiogonal method.

Using the coefficients of grid points, and by using the tridiogonal matrix algorithm,
the data distribution is calculated. The grid points are numbered 1,2,3,………….M.
with points 1 and M denoting extreme states.
The discretisation equation can be written as

Ai Ti + BiT i+1 +CiT i-1 = Di
For I = 1,2,3…M. Thus the data Ti is related to neighboring data values T i+1 and
T i-1. For the given problem C1=0 and BM=0 as T1 & TM represent boundary states.
These conditions say that T1 is known in terms of T2. The equation for I=2, is a
relation between T1, T2 & T3. But since T1 can be expressed in terms of T2 , this
relation reduces to a relation between T2 and T3. This process of substitution can be
continued until T M-1 can be formally expressed as TM. But since TM is known we can
obtain T M-1.This enables us to begin back substitution process in which T M-2,
T M-3…T3, T2 can be obtained. This process is continued until further iterations cease
to produce any significant change in the values of T’s. Finally the data distribution is
obtained for all grid points for different times by considering a suitable α which is
used as key.
3.3.4 Results
By considering a suitable key α =4, del t= 2, delx =2 for a total time stamp of 6 units,
Different data values obtained are
For del t=2, time =2;
33 6 7 4 33 8 11 13 32 22 29 20 26 0 18 10 17 11 1 1;
For delt =2, time=4;

8 22 4 3 5 11 11 13 5 30 22 4 17 14 28 27 29 29 15 1
3 30 2 6 27 12 10 15 29 1 26 26 3 32 0 4 18 8 1 32
For delt =2, time=6;

33 6 7 4 33 8 11 13 32 22 29 20 26 0 18 10 17 11 1 1;
3 26 34 17 16 29 11 19 0 23 22 11 33 6 14 13 3 1 4 7;
3 10 21 23 5 33 9 18 0 20 31 17 15 18 6 14 0 9 31 1;
Thus by using the same key, by changing the time stamp values different sequences
can be generated which are used as sub keys. These sub keys can be mapped to plain
text to generate cipher text.
Table No. 3.3.1: Encryption & Decryption process by Mathematical Model
Encryption
Plain Text A S K S
Conversion to 10 28 20 28
alpha numeric
value
Sub key 33 6 7 4
Total 43 34 27 32
Mod 36 07 34 27 32
Cipher Text 07 Y R W
Decryption
Cipher Text 07 Y R W
Conversion to 07 34 27 32
alpha numeric
value
Add 36 if less 43 34 27 32
than 9
Sub key 33 6 7 4
Subtract 10 28 20 28
Plain Text A S K S
3.3.5 Number of Computations for converting plain text to cipher text:
1. Considering steady state equation

2. Integrating the Steady state equation over control volume.
3. Representing the above equation in terms of a constant, time stamps and nonce.
4. Identifying the coefficients A[],B[],C[],D[] in terms of constant which is nothing
but key, time stamp & nonce.
5. Initialization of boundary data values says T1 & Tn.
6. Representing T2 in terms of T1 & T3 and continuing the process till Tn-1.
7. Known Tn , using back tracking to calculate Tn-1, Tn-2 ………………..T2.
8. Repeat the process for different steps of time stamps to generate sequence.
9. Mapping of sequence to Plain text to generate cipher text.
Thus the total number of computations to generate cipher text from plain text using the
sequence generated from the developed Model is 9.
Computing power to generate cipher text from plain text:

If we go by the algorithm, the number of computations required to generate the sequence
at different grid points by different time stamps depends on the steps of time stamp,
nonce considered. In the given problem for a time stamp 10 units & a Nonce of 21, the
total number of computation needed to generate 20 character block of cipher text is
around 500 computations.
3.3.6 Complexity of the model

By Construction: In the given model, the data value depends on a constant value which
is the key; it also depends on time stamp and a nonce value. By changing the time stamp
different values of the sequence which is to be mapped between plain text & cipher text
can be generated. By changing the nonce also, different values can be generated which
provides good security against crypto analysis. The complexity of the model by
construction is O(nt).
Table No. 3.3.2: Complexity of the Algorithm by its construction.
S.No. Computational Details No. of Computations Complexity
1 Read key, delt, delx 4
&Time
2 Read nonce(n) 1
3 Initialize grid points N O(n)
with a constant value
4 Initialize first T1 and 2
last Tn grid points
5 Calculate coefficients 4(n-2) O(n)
A[],B[],C[],D[] of n-2
grid points.for delt
6 Repeat the process for 4*(n-2)*(time/delt) O(nt)
successive delt till
time is reached
7 Calculate T2 in terms n-1 O(n)
of T1 & T3 and
continuing the process
till Tn-1 represented
in terms of Tn.
8 Known Tn calculate 1
Tn-1
9 Back tracking the n-1 O(n)
process to calculate
Tn-2 in terms of Tn-1
and so on till T2.
10 Repeat the process for n-1*(time/delt) O(nt)
successive delts till
total time is reached
Thus if we go by the algorithm, the complexity is O(nt).

By its Strength: Since in the given model present data depends on the key, time
stamps, nonce & Past data, there is no means by which the key can be guessed rather
than trying all combinations of keys. Thus the complexity of the model can be
considered exponential in nature.
3.3.7 Avalanche effect: It mainly deals with identifying major variations in cipher text
by small variations in key values. In the given model it is observed that a small change in
key values, Time stamps or the nonce is providing major changes in cipher values.
Case 1: By changing the key by one value.

α =5, del t= 2, delx =2, time =4; for n= 21.
Considering initialization constants T1 & Tn as 300 ,300
respectively.
12 8 11 19 30 17 7 14 22 9 4 16 28 26 22 9 25 7 10 6 0 12 27 1 35 22 7 15 23 2 19 7 15 8
25 20 31 14 6 2 25 0.
Case 2: α =5, del t= 1, delx =2, time =4;

12 6 20 5 9 4 28 22 32 17 5 28 7 29 28 34 22 32 26 20 0 12 31 22 3 24 7 5 23 30 12 27 19
22 0 11 35 15 6 15 30 0 12 12 17 23 1 6 17 3 19 27 11 15 14 24 29 13 6 18 19 28 22 0 12
21 12 4 8 5 25 9 2 21 15 1 25 31 2 1 16 1 18 7

12 10 9 10 9 10 9 10 9 10 9 10 9 11 6 16 5 6 24 13 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 4 6 7 0.

12 21 12 16 3 0 24 7 3 1 30 18 18 34 9 3 16 9 27 6 0 12 20 4 1 24 28 22 16 28 17 30 3 21
20 23 32 25 8 13 1 0.
12 6 4 25 26 20 19 22 9 29 13 32 9 33 19 23 32 15 23 12 2 12 26 29 15 30 18 16 28 30 13
16 13 28 26 5 16 12 1 9 6 2.
3.3.8 Security Analysis:

In the given model, the sub key generated depends on
a) Time stamp which is total time/del t.
b) The considered Nonce value, which gives variable number of sub key values.
Since both the time stamp and nonce considered can be dynamic in
nature, the number of rounds of the algorithm and the number of values of the sub key
generated are variable in nature. These features provide sufficient security to the
algorithm. Since the number of rounds is variable in nature and a mod function is applied
on the sub key values, it is relatively free from differential crypto analysis. The Nonce
value considered in the algorithm is dynamic which gives variable number of sub key
values which makes the algorithm free from linear crypto analysis.
This model is using dynamic time stamps and nonce values in addition to key values to
generate sub keys, it is relatively free from cipher text, known plain text, chosen plain
text, chosen cipher text and chosen text attacks
3.3.9 Conclusion:
This encryption mechanism uses a Initialization Vector, Time Stamp & Key to
generate distributed sequences which are used as sub-keys. Since the time stamp is
variable in nature, the model provides sufficient security against crypto analysis. The
model is free from cipher text attack, known plain text & cipher text attacks. In the
given model past & present time stamps have been used to generate data. By properly
guessing future time stamps, the model can be made still stronger.
Chapter 4
TRAINING OF THE DEVELOPED MODELS
4 Training of the Model
T raining an encryption model involves variable length key, a plain text and an
algorithm which applies the key on the plain text to generate cipher text. This
allows for the data to be transmitted over the network in some form which cannot be read
by any intruder.
In the first algorithm, the model will be trained for different keys and an analysis of
the models will be done. The keys that are used for training on the model are 3*3 keys
and 4*4 matrix keys. By having small variations in the keys, the model will be studied for
its increase in performance and working. The model will also be trained for its increase in
security by having slight variations in the key values. The model will also be trained for
changes in data overhead for small variations in the key values.
The model will be trained for different input vectors applied on the matrix keys to
generate different sequences which will be used to generate basins. The model is trained
for different input vectors of 27, 81 values with base 3 and 64 & 100 values with a base
4. For example, if the input vector of 27 values ranging from(0-26) is considered, with a
base value of 3 & the number being represented in 3 digit form, then the values will be
represented as (0 0 0 to 2 2 2).
For example,
(26) 3 in 3 digit format = 2 *3 2 +2*3 1 +2*3 0.= ternary equivalent is ( 2 2 2) 3
(80) 3 in 4 digit format = 3*3 3+3*3 2+3*3 1+3*3 0 = ternary equivalent is (3 3 3 3) 3
(100) 4 in 4 digit format = 1*4 3+2*4 2+1*4 1+0*4 0. = 4-ary equivalent is (1 2 1 4) 4
(255) 4 in 4 digit format = 3*4 3+3*4 2+3*4 1+3*4 0.= 4-ary equivalent is ( 3 3 3 3) 4
For example,
Case 1: Considering the key A of size (3*3)
2 5 −6
A = key 3 1 3
4 − 2 −3
n=0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
26
r=2 0 0 18 0 3 18 18 6 20 2 6 20 13 6 20 24 6 20
r= 8 8 23 26 8 26 26 24.
the basins that can be formed using this sequences are

b(0)=(0,1,2,4,10)
b(1)=(3,5,6,7,8,9,11,12,14,15,17,18,19,20,21,23)
b(2)=(13)
b(3)=(16,22,24,25,26)

3 5 −6
A = key = 4 1 3
5 − 2 −3
Sequence generated
r= 1 0 0 18 0 0 18 18 3 20 2 6 20 13 6 20 24 6 23 8 8 26 26 8 26 26 25.
Thus the basins formed are

b(0)= (0,1,2,4,5,10)
b(1)= (3,6,7,8,9,11,12,14,15,17,18,19,20,23)
b(2)=(13)
b(3)= (16,21,22,24,25,26);
Case 3:
1 5 −6
A = key = 2 1 3
2 − 2 −3
r= 11,1,3,20,0,6,18,18,6,20,2,6,20,13,6,20,24,6,20,8,8,20,26,6,23,25,15.
Basins formed are
b(0) = (0,11,4)
b(1) = (1);
b(2)= (2,3,10);
b(3)= (5,6,7,8,9,11,12,14,15,17,18,19,20,21,22,23,24,26)
b(4)=(13)
b(5)=(25).
In the case 3, for the given key, 6 basins have been formed. By this key, the alphabet
supports 6 characters. By properly choosing each alphabet value by two basins, the
number of characters of the alphabet can be increased to 36 values. This reduces the data
overhead by half of its value.
Case 4:
By increasing the key size to [4*4], we can increase the number of basins and also the
values of each basin, which increases the strength of the algorithm.
1 5 −6 1
2 1 3 2
A = key =
3 − 2 −3 3
4 2 4 4
Sequence generated for n=0 to 80 values represented to the base 3.
r=0,33,60,0,0,6,0,9,20,54,57,60,0,0,40,0,19,20,54,54,60,54,54,74,9,20,20,33,60,60,0,6,26
,9,20,26,57,60,61,0,40,80,19,20,23,54,60,71,54,74,80,20,20,47,60,60,71,6,26,26,20,26,26
,60,61,80,40,80,80,20,23,26,60,71,80,74,80,80,20,47,80.
The basins formed by considering similar values,
b(0)=(0,3,4,6,12,13,15,30,39,5,31,57,10,36)
b(1)= (9,54,7,24,33,18,19,21,22,45,48,1,27,16,42)
b(2)= (20,60,8,17,25,26,34,43,51,52,69,78,2,11,28,29,37,46,54,55,63,72,32,35,
58,59,61,62,71,9,18,19,21,22,45,48,38,64,47,56,73,7,24,33,16,42,53,79,1,27)
b(3)= (23,74,44,70,49,75)
b(4)= (40,14,66)
b(5)= (80,41,50,65,67,68,76,77).
Case 5:
2 5 −6 1
3 1 3 2
A= key =
4 − 2 −3 3
5 2 4 4
Sequence generated n=0 to 80 values represented to the base 3.

r=0,6,33,0,0,6,0,0,20,54,54,60,0,0,0,0,9.20,54,54,57,54,54,65,0,20,20,33,60,60,0,6,26,9,2
0,26,57,60,61,0,40,80,19,20,23,24,60,71,54,74,80,20,20,47,60,60,80,15,26,26,23,26,26,6
0,71,80,80,80,80,20,26,26,26,16,80,80,74,80,80,47,74,80.
The basins formed from the generated sequence
b(0)=(0,3,4,6,7,12,13,14,15,24,30,39,1,5,31,57,45,20,36,8,17,25,26,34,43,44,51,52,69,32
,35,58,59,61,62,70,71,38,47,64,53,78)
b(1)= (9,54,16,33,10,18,19,21,27,48,72,2,42)
b(2)= (23,65,60,11,28,29,37,46,54, 55,63,9,10,18,19,21,22,48,16,33,42,72,2,27)
b(3)= (40)
Case 6:
3 5 −6 1
4 1 3 2
A = key =
5 − 2 −3 3
6 2 4 4
Sequence generated for n=0 to 80 values being represented to the base 3.

r=0,3,6,0,0,3,0,0,10,54,54,60,0,0,0,0,0,20,54,54,54,54,54,55,0,10,20,33,60,60,0,6,26,9,20
,26,57,60,61,0,40,80,19,20,23,54,60,71,74,54,80,20,20,47,60,70,80,25,26,26,26,26,26,60,
80,80,80,80,80,20,26,70,80,80,77,80,80,74,77,80.
Basins formed from the generated sequence
b(0)=(0,3,4,6,7,12,13,14,15,16,22,28,37,79,80,1,5,2,29,39,48,54,62,63,64,65,66,67,71,72
,74,75,78,9,10,18,19,20,43,47,45,46,76,31,8,23,40,17,24,32,41,49,50,68,51,42,38)
b(1)=(25,33,55,21,26,60,30,56,57,58,59,69,11,27,35,44,52,61,34,36)
b(2)= (70,53)
b(3)= (77,73)
Case 7:
1 5 −6 1
2 1 3 2
A = key =
3 − 2 −3 3
4 2 4 4
Sequence generated for n=0-100 for a base value of 4.
r=0,72,136,136,0,0,8,42,0,16,34,42,32,34,34,34,128,132,136,137,0,0,85,170,0,33,34,38,3
3,34,34,34,128,128,136,154,128,128,162,170,16,34,34,98,34,34,34,34,128,128,170,128,1
45,162,162,161,162,162,162,34,34,34,34,72,136,156,154,0,8,42,43,16,34,42,42,34,34,34,
42,132,136,137,170,0,85,170,33,34,38,42,34,34,34,38,128,136,154,170,128.
Basins formed the generated sequence
b(0)= (0,4,5,8,20,24,21,68,84,6,69)
b(1)=(16,128,9,40,72,32,33,36,37,48,49,52,96,100,1,64,12,25,28,38)
b(2)=(34,136,10,13,14,15,26,29,30,31,41,42,44,45,46,47,60,61,62,63,73,76,77,78,89,92,
93,94,2,3,18,65,81,97,7,14,70,71,74,75,79,91)
b(3)=(38,162,27,90,95,54,55,57,58,59,)
b(4)=(85,22)
b(5)=(98,154,43,35,67)
Case 8:
2 5 −6 1
3 1 3 2
A = key =
4 − 2 −3 1
5 2 4 4
Sequence generated for n=0:100 for a base value of 4.
r=0,8,72,136,0,0,8,26,0,0,34,42,16,34,34,128,128,136,136,0,0,0,106,0,16,34,34,32,34,34,
34,128,128,132,138,128,128,146,166,0,34,34,34,34,34,34,34,128,128,128,154,128,128,1
62,162,144,162,162,162,34,34,34,34,72,136,136,154,0,8,42,42,16,34,42,92,34,34,34,42,1
32,136,137,170,0,85,170,170,33,34,38,42,34,34,34,38,128,136,154,170,128.
Basins formed from generated sequence
b(0)=(0,45,8,9,20,21,22,24,40,68,84,1,6,69)
b(1)=(16,128,12,25,72,17,32,33,36,37,48,49,50,52,53,96,100,2,64,28,88)
b(2)=(26,34,7,10,13,14,15,27,29,30,31,41,42,43,44,45,46,47,60,61,62,63,73,76,77,78,89,
92,93,94,11,70,71,74,79,91,75)
b(3)=(38,146,90,95)
b(4)=(85)
3 5 −6 1
4 1 3 2
Case 9: A = key =
5 − 2 −3 3
6 2 4 4
Sequence generated for n=0:100 for a base value of 4.
r=0,4,8,72,0,0,4,9,0,0,17,38,0,33,34,34,128,128,136,136,0,0,26,0,0,34,34,16,34,34,34,12
8,128,128,137,128,128,129,162,17,34,34,33,34,34,34,128,128,128,134,128,128,146,162,
64,162,162,162,34,34,34,72,136,156,154,0,8,42,42,16,34,42,42,34,34,34,42,132,136,137,
170,0,85,170,170,33,34,38,42,34,34,34,38,128,136,154,170,128
Basins form the generated sequence
b(0)=(0,4,5,8,9,12,20,22,23,38,65,80,97,98,99,100,1,6,2,66,7,11,86,91)
b(1)=(16,128,26,69,15,30,31,32,34,35,46,47,48,50,51,92,96,21,14,24,25,27,28,29,40,41,
43,44,45,5,56,60,70,73,74,75,85,88,89,90,81)
b(2)=(17,136,10,39,18,62,63,78,93)
b(3)=(33,137,13,42,84,67,68,71,72,76,87,3,61)
b(4)=(64,154,54,94)
From the above study, it has been observed that the generated model is trained for a [3*3]
and [4*4} random keys on an input ternary vector of 27 & 81 values with a base of 3, and
on 100 values with a base of 4. An attempt has been made to show better a avalanche
effect with respect to the models developed. For the input vectors of 27, 81 & 100 values,
the number of basins formed varies between 3to6. It has been observed that the more the
values of the vector, it does not guarantee in the increase of basins. But increases in the
values of basins are observed which provides more security. It is also observed that for a
certain variations in the key, the model is generating more basins. Thus for a random
generated key, the number of basins is evaluated. A better metric can be provided to a
random key which provides more number of basins. Since the numbers of basins form the
characters of the alphabet, the more the number of basin, and the more will be the
characters of the alphabet. Thus a combination of basins will increase the number of
characters of the alphabet. It is also observed that the more the number of basins, the less
will be the data overhead. For example, if the basins generated are 6 and if 2 basins are
considered for each character, and then the number of characters of the alphabet will be
36. If the number of basins formed is only 3, then each character has to be represented by
3 basin values if the alphabet is supporting 27 characters. The analysis and results of the
Training model is illustrated in the table no. 4.1 & 4.2.
Table 4.1: Relationship between Random Key considered with the Basins ( Sub
Keys) generated
S.No. Cases: Ternary Random Key Number of Metric
vector of a 3 & 4 Considered Basins between
digit number formed Key and
Basins
formed
1 Considered to the [2 5 -6; 3 1 3; 4 -2 - 4 Low
base 3 3]
2 Considered to the [3 5 -6; 4 1 3; 5 -2 - 4 Low

base 3 3]
3 Considered to the [1 5 -6; 2 1 3; 2 -2 - 6 High
base 3 3]
4 Considered to the [1 5 -6 1;2 1 3 2;3 -2 6 High
base 3 3 3;4 2 4 4]
5 Considered to the [2 5 -6 1;3 1 3 2;4 -2 4 Low
base 3 -3 3;5 2 4 4]
6 Considered to the [3 5 -6 1;4 1 3 2;5 -2 4 Low
base 3 -3 3;6 2 4 4]
7 4- ary vector of a [1 5 -6 1;2 1 3 2;3 -2 6 High
4 digit number -3 1;4 2 4 4]
considered to the
base 4
8 Considered to the [2 5 -6 1;3 1 3 2;4 -2 5 Medium
base 4 -3 1;5 2 4 4]
9 Considered to the [3 5 -6 1;4 1 3 2;5 -2 5 Medium
base 4 -3 3;6 2 4 4]
Table 4.2: Level of Metric between Key and Basins (Sub Keys) generated for Sign
Function applied on the product of Modified ternary vector & the random matrix
key
S Random Key No. of Ratio of Number of Data Metric that

no. Considered basins Cipher characters overhead can be
formed text to of alphabet mapped
plain text between Key
and Basins
formed
1 [2 5 -6; 3 1 3; 4 3 64 3 Low
4 -2 -3]
2 [3 5 -6; 4 1 3; 4 3 64 3 Low
5 -2 -3]
3 [1 5 -6; 2 1 3; 6 2 36 2 High
2 -2 -3]
4 [1 5 -6 1;2 1 3 6 2 36 2 High
2;3 -2 3 3;4 2
4 4]
5 [2 5 -6 1;3 1 3 4 3 64 3 Low
2;4 -2 -3 3;5
2 4 4]
6 [3 5 -6 1;4 1 3 4 3 64 3 Low
2;5 -2 -3 3;6
2 4 4]
7 [1 5 -6 1;2 1 3 6 2 36 2 High
2;3 -2 -3 1;4
2 4 4]
8 [2 5 -6 1;3 1 3 5 3 125 3 Medium
2;4 -2 -3 1;5
2 4 4]
9 [3 5 -6 1;4 1 3 5 3 125 3 Medium
2;5 -2 -3 3;6
2 4 4]
Table 4.3: Level of Metric between Key and Basins (Sub Keys) generated for
different functions applied on the product of Modified ternary vector & random key
S no. Random Function used No. of Metric that can be mapped

Key in the model on basins between Key and Basins
Considered the product of formed formed
ternary vector
& random key
1 [0 2 3;2 2 Sign 3 Low
2;0 2 2]
2 [9 8 9;10 Sign 3 Low

13 10;9 9
9]
3 [-2 -2 -3;-4 Sign 2 Low
-2 -2;-2 -2 -
2]
4 [10 -8 -9;- Sign 4 Low
11 13 10;10
9 9]
5 [1 5 -6; 2 1 Sign 6 High
3; 2 -2 -3]
5 [1 2 3;3 2 Mod, 3 3 Low
2;
1 2 2]
6 [1 2 3;3 2 Mod, 5 6 High
2;
1 2 2]
7 [1 2 3;3 2 Mod, 7 6 High
2;
1 2 2]
8 [11 8 9;12 Mod, 3 8 High
13 10;11 9
9]
9 [11 8 9;12 Mod, 5 7 High
13 10;11 9
9]
It is observed that the by using different functions like Modular arithmetic or Sign
functions on the product of Modified ternary vector and matrix key, the relation to key
and sub keys almost remains the same ie varies between 3 -8 basins(sub keys).
Table 4.4 : Possible Relationship between Plain text and Cipher Text Generated in
Probabilistic encryption Algorithm (Model 3) (Depending on the random Key
Chosen).
S no. Ratio of Number of Possible Number of Possible

Cipher text basins characters values in each Conversions
to Plain text formed in the Basin of one plain
alphabet text character
to one cipher
text character
Case 1 3 3, if single 27 6,14,5 30-450
valued basin
is neglected.
Case 2 3 4 64 6,1,13,6 6-78
Case 3 2 6 36 3,1,3,1,1,17 3-51
Thus depending on the key considered and the number of sub keys(Basins) generated,
one plain text character can be replaced by different combinations of cipher texts. As per
the above tables it is observed that if one plain text character is replaced by 3 basins, then
the number of possible cipher text characters for one plain text character is 30 to 450
combinations. For different possible keys different combinations of cipher texts are
possible for one plain text character which varies from a lesser value to higher values.
Conversion of plain text to cipher text

Table 4.5: Conversion of plain text to cipher text for a 5 character text
Case 1:
Plain Text for a 5 Character text The re
Cipher text for model 1 Vhesm
Cipher text for model 2 Wmkym
Cipher text for model 3 Vbp0vvjspvdj0ex
Case 2:
Plain Text for a 10 Character text The require
Cipher text for model 1 Vhesmqxjzf
Cipher text for model 2 Wmkymrvjjf
Cipher text for model 3 Xapapy0izyoedkveypxudqoayabdix
Case 3:
Plain Text for a 15 Character text The requirements
Cipher text for model 1 Vhesmqxjzfukptu
Cipher text for model 2 Wmkymqfsuz
Cipher text for model 3 Ydp0ppmbby00dtpnpyzo0sddvdcjsychrdsyrgyp0vxai
Case 4:
Plain Text for a The reqiurementsof inf
20 Character
text
Cipher text for Vhesmqxjzfukptuuhioi
model 1
Cipher text for Wmkymrvjjfqfsuzpnjwh
model 2
Cipher text for V0pdvyacpybal0yspxzkdlday00dqycchdoyelpyappdolpdbydqaasrxdxa
model 3
Table 4.9: Multiple Cipher Texts to one Plain Text

In a probabilistic encryption schemes (Model3) we can see that for the same plain text,
multiple cipher texts can be generated.
Plain text for a 5 character value The re
Cipher text 1 Vbp0vvjspvdj0ex
Cipher text 2 Xapapy0izyoedkv
Cipher text 3 Ydp0ppmbby00dt
Cipher text 4 V0pdvyacpybal0y
Thus multiple cipher texts can be formed for one plain text in probabilistic
encryption scheme which makes the algorithm strong from chosen cipher text attacks.
Training of second Algorithm:
In the Second algorithm, the model will be trained for different keys, time stamps &
Nonce values and an analysis of the model will be done. In the given model it is observed
that a small change in key values, Time stamps or the nonce is providing major changes
in cipher values.
Case 1: α =5, del t= 2, delx =2, time =4; for n= 21.

Considering initialization constants T1 & Tn as 300 ,300
respectively.
12 8 11 19 30 17 7 14 22 9 4 16 28 26 22 9 25 7 10 6 0 12 27 1 35 22 7 15 23 2 19 7 15 8
25 20 31 14 6 2 25 0.

12 6 20 5 9 4 28 22 32 17 5 28 7 29 28 34 22 32 26 20 0 12 31 22 3 24 7 5 23 30 12 27 19
22 0 11 35 15 6 15 30 0 12 12 17 23 1 6 17 3 19 27 11 15 14 24 29 13 6 18 19 28 22 0 12
21 12 4 8 5 25 9 2 21 15 1 25 31 2 1 16 1 18 7

12 10 9 10 9 10 9 10 9 10 9 10 9 11 6 16 5 6 24 13 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 4 6 7 0.

12 21 12 16 3 0 24 7 3 1 30 18 18 34 9 3 16 9 27 6 0 12 20 4 1 24 28 22 16 28 17 30 3 21
20 23 32 25 8 13 1 0.

12 6 4 25 26 20 19 22 9 29 13 32 9 33 19 23 32 15 23 12 2 12 26 29 15 30 18 16 28 30 13
16 13 28 26 5 16 12 1 9 6 2.
Thus we can see that by changing the key values slightly, by changing the time stamp the
model is generating variable and distinct values which provides sufficient strength to the
algorithm. The algorithm can be made still stronger by varying the nonce value which
generates variable sub key values.
4.2 Role of Statistical tests on Values generated by the algorithms under study.
In the given study, two algorithms are discussed. Both the algorithms execute a series of
steps and generate a sequence. This sequence is being used as sub key to be mapped to
plain text to generate cipher text. The strength of the encryption & Decryption process
depends on the strength of sequence generated. In this part of work some statistical tests
like Uniformity tests, Universal tests & Repetition tests are tried on the sequence
generated to test the strength of it.
Uniform test tries to study the uniformity of a sequence generated. From the sequence
generated, the first two consecutive points are considered as coordinates of a graph. The
process is repeated for all the values of the sequence to generate the graph. For example,
if a,b,c,d……..n are the values of sequence generated, then the coordinates of the graph
are (a,b),(c,d)……….(n-1,n). If they are dependent they will form into lines or they will
form into a plane. In this study, different cases are tried for both the algorithms to study
the non uniformity distribution of values in the sequence.
Universal tests: It tries to see whether the data can be compressed. Since in the given
sequence, the repetition of values is minimum, we can say that the developed sequences
are relatively free from this test.
Repetition test: This test studies that each character of the sequence is not identical to the
earlier character. In the sequence generated by the first algorithm, the values are random
and unique in nature, the probability that the values get repeated is 1/27 . In the second
algorithm even though some values of the sequence are repeated, the distribution of
values is not uniform and random in nature which does not give any insight about he
pattern of the sequence. Thus we can say that the sequence generated by both the
algorithms is relatively free from Uniformity test, Universal test & Repetition tests.
The analysis is being done on sequences generated by both the algorithms which are
represented as case studies in the earlier part of the work. The analysis is being
represented as graphs for both the algorithms.

Fig 4.1: Sequential Data Representation of Algorithm 1:
30 30
25 25
20 20
Scatterd Scattered
15 values for case 15 values of case
1 2
10 10
5 5
0 0
0 10 20 30 0 10 20 30
30 90
80
25
70
20 60
Scattered Scattered
50
15 values of values of case
40
Case 3 4
10 30
20
5
10
0 0
0 10 20 30 0 50 100
90 90
80 80
70 70
60 60
Scattered Scattered
50 50
values of case values of case
40 40
5 6
30 30
20 20
10 10
0 0
0 50 100 0 50 100
Fig 4.2: Statistical Representation of Data of Algorithm 2:
30 35
25 30
25
20
sequential 20 Sequential
15 values of alg data of alg 2
2: Case 1 15 Case 2
10
10
5 5
0 0
0 20 40 0 10 20 30
18 40
16 35
14 30
12 25
10 Sequential Sequential
values of alg 20 values of alg 2
8
2: case 3 15 Case 4
6
4 10
2 5
0 0
0 10 20 30 0 20 40
35
30
25
20 Sequential
values of alg
15 2: Case 5
10
5
0
0 20 40
From the graphs we can say that the data distribution is non uniform and random in
nature in the sequence generated by the algorithms which provides sufficient strength
against crypto analysis.
Chapter 5
Key Distribution Mechanism
Key distribution mechanism forms an integral part of encryption decryption process. The
success of Symmetric encryption lies with proper and confidential sharing of private key
among the participating parties. Some times keys get garbled in transmission. Since a
garbled key can mean mega bytes of unacceptable cipher text, this sis a problem. All keys
should be transmitted with some kind of error detection and correction bits. In this part of
work, a new approach is tried to identify any garbled form of key is transmitted to
participating parties. A hash value is generated from the key to be transmitted. The
generated hash value which is appended to the key encrypted by either private keys of
the participating parties or the master key shared by the participating parties provide
security & confidentiality to key transmitted and also identifies any garbling of key
values during transmission.
In this part of work, a new approach is tried to identify whether any garbled form of key
is transmitted to participating parties.
In the methodology of the proposed work, two algorithms are presented, the first
algorithm deals with generating the product of ternary vector and a random matrix key
and applying a sign function on the product to generate a sequence (pp 42). In the second
algorithm a mathematical model is considered which uses a key, time stamp and nonce
values to generate sequence (pp 84).
5.1 Model for generation of Hash value for given key.
1. Generated sequence by the earlier methods is considered.

2. The sequence is converted to binary form.
3. A block of bits of binary sequence say 6 is considered.
4. Each block is exclusive_or’ ed with the next block of data to generate a 6 bit
hash value.
5.2 Example:
Considering the first algorithm,
1. Generated sequence from the algorithm 0 13 11 6 2 12 10 5 1
2. Binary form of the sequence= 00000000 00001101 00001001 00000110 00000010
00001100 00001010 00000101 00000001.
3. Considering 6 bits as one block of data.
4. Generated hash value= 101111 = 47.
5.3 Identifying any garbled key during transmission:

Appending the hash value to key: When the key is being transmitted to the participating
parties, a hash value is generated which will be appended to the key. At the receiver’s side,
considering the key and using the same algorithm hash value is regenerated. This
regenerated hash value is compared with received hash value. If both the hash values are
matching, it means that key received is in correct fashion and no garbling of keys has taken
during transmission. The advantage with this mechanism is, if any errors takes place in the
key while transmission, there will be a mismatch between received hash value & generated
hash value and accordingly the participating parties ask for retransmission of key.
5.4 Conclusion:
This mechanism uses the same algorithm for the generation of sub keys and also for the
generation of sequence which is used to generate hash value. If any errors takes place in the
key while transmission, there will be a mismatch between received hash value & generated
hash value and accordingly the participating parties ask for retransmission of key. Thus this
model identifies any garbled form of keys while transmission from Key Distribution
Centre.
Chapter 6
COMPARATIVE STUDY OF DIFFERENT
DEVELOPED MODELS
Comparative Study of Different models of the proposed work.

The core of the study is the analysis and discussion of Developed encryption
models. All the models are studied for their performance in terms of computational
power, avalanche effect, complexity of the models in terms of their construction and
strength. The models are also studied for their improved performance against crypto
analysis.
6.1 Number of Computations for the developed models

Algorithm1:
Model 1: 10 Computations to convert one plain text character to one cipher text
character.
character.
character.
Algorithm 2:
Model : 09 Computations to convert one plain text character to one cipher text
character.
Computational overhead (Computing Power) for the developed models
Algorithm 1:
Model 1:
Computation overhead for a 9 character key
1st computation: 27 calculations, 2 computation: 27 calculations, 3 computation: 27*9
calculations. 4th computation: 27 calculations, 5th computation: 27 calculations, 6th
computation: 27 calculations, 7 computation: 27 calculations considering a 27 character
plain text, 8th computation: 27 calculations, 9th computation: 27 calculations, 10th
computation 27 calculations.
Thus the total computational overhead per block of data by the first model is 486
calculations.
Model 2:
Computational overhead for a 9 character key
1st computation: 27 calculations, 2nd computation: 27 calculations, 3rd computation:
27*9 calculations, 4th computation: 27 calculations, 5th computation: 27 calculations,
6th computation: 27 calculations, 7th computation: 27*27 calculations +27*27
calculations, 8th computation: 27 calculations, 9th computation: 03 calculations
depending on chosen rule, 10 computation 03 calculations, 11 computation: 27
calculations, 12th computation: 27 calculations, 13th computation: 27 calculations, 14th
computation: 03 calculations depending on the chosen rule.
Thus the total computational overhead per block of data by the second model is 1953
calculations.
Model 3:
Computational overhead for a 9 character key
1st computation: 27 calculations, 2nd computation: 27 calculations, 3rd computation:
27*9 calculations, 4th computation: 27 calculations, 5th computation: 27 calculations, 6th
computation: 27 calculations, 7th computation: 27*27 calculations +27*27 calculations,
8th computation: 27 calculations, 9th computation: 81 calculations, 10th computation:
81 calculations, 11th computation: 81*k calculations (where k refers to number of
computations for a random pick up of values). Thus the total computational overhead per
character by the fourth model is 2025 +81k calculations.
Algorithm 2:
If we go by the algorithm, the number of computations required to generate the sequence
at different grid points by different time stamps depends on the steps of time stamp,
nonce considered. In the given problem for a time stamp 10 units & a Nonce of 21, the
total number of computation needed to generate 20 character block of cipher text is
around 500 computations.
Table 6.1:Comparative Study of Number of Computations and computational

overhead of generated models.
Algorithm Number of computations Computational overhead for

for one plain text character one plain text character to
to one cipher text character. one cipher text character.
Algorithm 1: Model1 10 486
Algorithm 1: Model2 14 1953
Algorithm 1 Model3 11 2025+81k
Algorithm 2: Model 09 500
6.2 Complexity of the Model by Construction & Strength.
Algorithm 1:
Model 1:
Complexity of the Model by construction
If we go by the construction of the algorithm, there are 10 computations in the algorithm.
All the computations are product of the values of ternary vector. If n is the number of
values of ternary vector considered, we can say that the complexity of model1 is O (n).
Complexity of the Model by its strength
In the given algorithm a matrix key is used. This matrix key is multiplied with ternary
vector. On the generated values a sign function is used to convert all positive values to 1,
negative values to –1 and zero to 0. This provides the necessary strength to the algorithm.
Thus known the algorithm, known the cipher text it is computationally infeasible to
generate the matrix key. Thus if we go by the security of the algorithm, the complexity is
exponential in nature.
Model 2:
All the computations are product of the values of ternary vector. For the basins generated
computation, it is the square of the values of the ternary vector. If n is the number of
values of ternary vector considered, we can say that the complexity of model2 is O (n 2).
Complexity of the Model by Strength
negative values to –1 and zero to 0. A sequence is generated. Based on the similarity of
values this sequence is divided to basins, These basins with different values are mapped
to plain text to generate cipher text. Thus this algorithm provides different keys with
variable length. Thus this algorithm is supposed to be safe in real time environment. Thus
if we go by the security of the algorithm, the complexity is exponential in nature.
Model 3:
All the computations are product of the values of ternary vector. For the basins generated
computation, it is the square of the values of the ternary vector. If n is the number of
values of ternary vector considered, we can say that the complexity of model3 is O (n 2).
Complexity of the Model by its strength
negative values to –1 and zero to 0. A sequence is generated. Based on the similarity of
values this sequence is divided to basins. Basins will form the number of characters of the
alphabet. A combination of basins are used which increases the number of characters of
the alphabet. Values from the basins are mapped to the values of plain text in a random
fashion to form cipher text. Thus if we go by the security of the algorithm, the complexity
is exponential in nature.
Algorithm 2:
Complexity of the model
By Construction: In the given model, the data value depends on a constant value which
is the key; it also depends on time stamp and a nonce value. By changing the time stamp
different values of the sequence which is to be mapped between plain text & cipher text
can be generated. By changing the nonce also, different values can be generated which
provides good security against crypto analysis. The complexity of the model by
construction is O(nt).
By its Strength: Since in the given model present data depends on the key, time
stamps, nonce & Past data, there is no means by which the key can be guessed rather than
trying all combinations of keys. Thus the complexity of the model can be considered
exponential in nature.
TABLE 6.2: Comparative study of Different Generated Models in terms of Security
analysis by construction and crypto Analytical Strength
Model No. Security analysis by Crypto Analytical Strength

construction
Algorithm 1:1Model 1 O(n) Exponential
Algorithm 1: Model 2 O(n 2) Exponential
Algorithm 1: Model 3 O(n 2) Exponential
Algorithm 2: Model 1 O(nt) Exponential
6.3 Avalanche effect:

Algorithm 1:
Model 1:
By changing the key slightly, there is a lot of variations in the sequence generated
which provides maximum avalanche effect to the algorithm. Thus this model provides for
good variations in cipher text for slight variations in the key. This provides for maximum
strength and security to the algorithm. But since the model is a simple substitution
algorithm, the plain text variations to cipher text variations is negligible.
Model 2:
By changing the key slightly, there is a lot of variations in the basins generated.
Since these basins will form the keys for the given model, they provide maximum
avalanche effect to the algorithm. Thus this model provides for good variations in cipher
text for slight variations in the key. Depending on the starting value of the plain text, a
slight variation in plain text will call for a different key. Thus we will also see a lot of
variations in the cipher text generated for slight variations in the plain text. Thus it
provides a maximum avalanche effect to the algorithm which provides for more strength
and security.
Model 3:
For the same plain text, different cipher texts will be formed which makes crypto
analysis difficult. And also a lot of variation in the cipher text is observed by slight
variations in the key and the plain text. This provides for maximum avalanche effect on
the algorithm And also it is observed that by slight variations in the key, there is a
substantial variation in the number of basins formed, which again provides for better
avalanche effect which ultimately provides a maximum strength and security to the
algorithm.
Algorithm 2:
In the given model it is observed that a small change in key values, Time stamps or the
nonce is providing major changes in cipher values.
6.4 Differential and linear crypto analysis:

The overall strategy of differential crypto analysis that could be applied on DES
algorithm is based on the following considerations for a single round. The procedure is to
begin with two plain text messages m& m’ with a given difference and trace through a
probable pattern of difference after each round to yield a probable difference for the
cipher text. Actually these two probable differences are for the two 32 bit halves. Next
the plain texts m & m’ are considered for encryption to determine the actual difference
under the unknown key and the result of probable difference is measured. If there is a
match, that all probable patterns considered may be correct. With that assumption some
deductions about the key bits will be made. This procedure must be repeated many times
to determine all the key bits.
The objective of linear crypto analysis is to find a effective linear equation of the
form P[x1….xn] ex or C[y1…….ym]=K[z1…..zl]. where P & C refers to plain and
cipher texts and K be the key. Once a proposed relation is determined, the procedure is to
compute the results of the left-hand side of the equation for a large number of plain text
& cipher text pairs. If the result is zero for more than half of the time, assume that
K[z1…zl] = 0 other wise it is assumed as 1. This gives us linear equations on key bits.
These equations can be solved to generate key bits.
Algorithm 1:
Model 1:
cipher texts. Thus this model is free from differential crypto analysis.
But this model uses a simple substitution technique to generate cipher text; it is
some what susceptible to linear crypto analysis. The Key can not be gained and not a
whole of information can be gained, but part of information can be gained in this model.
Model 2:
Since in this work, it uses a sign function to generate the sequence. Similarity of
values of the sequence will be used to generate basins which are used as variable length
keys to map plain text to cipher text. Since it uses a sign function, it is free from
differential crypto analysis and since variable length keys are used it is also free from
linear crypto analysis.
Model 3:
Since in this work, it uses a sign function to generate the sequence. Similarity of
values of the sequence will be used to generate basins. These basins are mapped to
characters of plain text. Then the values from basins are chosen randomly to generate
cipher text. Since it uses a sign function, it is free from differential crypto analysis and
since basins with different and variable values are mapped with characters of plain text in
a random way to form cipher text, it is also free from linear crypto analysis.
Algorithm 2:
The time stamp and nonce considered can be dynamic in nature, the number of rounds of
the algorithm and the number of values of the sub key generated is variable in nature.
These features provide sufficient security to the algorithm. Since the number of rounds is
variable in nature and a mod function is applied on the sub key values, it is relatively free
from differential crypto analysis. The Nonce value considered in the algorithm is
dynamic which gives variable number of sub key values which makes the algorithm free
from linear crypto analysis.
6.5 Security Analysis :

The process of attempting to discover the plain text or key or both is known as
cryptanalysis. The various types of cryptanalytic attacks based on the amount of
information known are 1. Cipher text only where Encryption algorithm and cipher text to
be decoded are known. 2. Known plain text where encryption algorithm, cipher text and
one or more plain text cipher text pairs formed with the secret key are known, 3. Chosen
plain text where encryption algorithm, cipher text and plain text message chosen by
cryptanalyst together with its corresponding cipher text generated with the secret key are
known. 4. Chosen cipher text where encryption algorithm, cipher text and purported
cipher text chosen by the cryptanalyst together with the corresponding decrypted plain
text generated with the secret key are known. In the present work all the different models
developed by the proposed work are studied in terms of cryptanalysis.
Algorithm 1:
Model 1: This model is completely free from cipher text only, type of attack. By the
other attacks, the key may not be retrieved but a part of plain text may be retrieved.
Model 2: This model is completely free from cipher text only and known plain text
attacks but may not be completely free from chosen plain text and cipher text attacks
where a part of information may be leaked out by cryptanalysis.
Model 3: This model is completely free from cipher text, known plain text, chosen
plain text, chosen cipher text and chosen text attacks. This algorithm is also free from
public key attacks.
Algorithm 2:
This model is using dynamic time stamps and nonce values in addition to key
values to generate sub keys, it is relatively free from cipher text, known plain text, chosen
plain text, chosen cipher text and chosen text attacks
Table 6.3: Comparative Study of Different Generated Models by Avalanche effect,
Linear & Differential Crypto Analysis & Security Analysis
Model No. Avalanche Avalanche Effect by Effect by Security

effect by effect by Differential Linear Analysis
variations in variations in Crypto Crypto
key Plain text Analysis Analysis
Algorithm1 11 1 Low High Cipher text
Model 1:
Model 2 12 3 Low Low Cipher text &
Known Plain
Text
Model 3 Indeterminate Indeterminate Low Low Cipher text,
Known Plain
Text, chosen
plain text &
cipher text.
Algorithm 20 1 Low Low Cipher text,

2 Known Plain
Text, chosen
plain text &
cipher text.
Table 6.4: Summary of Comparative study of some models of the proposed
algorithm in terms of number of computations, computational overhead, change in
plain text , change in key, complexity of the models by their construction and
strength and security analysis, depending on the chosen key.
Mod Type of No. of Computat Change Change Complexi Complexi Security

el no. Cipher computati ional in plain in key ty by its ty by its analysis
ons overhead text constructi strength
per on
character
Algor Block 10 486 1 11 O(n) Exponenti Cipher
ithm Cipher al text only
1:Mo
del 1
Mode Block 15 1953 3 12 O(n**2) Exponenti Cipher
l2 Cipher al text,
known
plain text
Mode Stream 11 2025+81k Indetermi Indetermi O(n**2) Exponenti Text only
l3 Cipher nate nate al
Algor Stream 09 500 1 20 O(nt) Exponenti Text only

ithm Cipher al
2
Fig.6.1 Pictorial Representation of Comparative study of some models of the
proposed algorithms in terms of computational overhead, change in plain text ,
change in key Depending on the chosen Key.
2500
2000
1500
1000 Computational
Overhead
500
0
Model 1 Model 3
20
15
10
Changes in Key
0
Model 1 Model 2 Model 3 Algorithm 2
2.5
1.5
Changes in Plain text
1
0.5
0
Model 1 Model 2 Model 3 Algorithm 2
Table 6.5.1: Comparative study of the proposed Block Ciphers( Model 1 & Model2)
with standard algorithm like DES in terms of computational overhead, Data
overhead, complexity and security analysis.
Algorithm Computational Data Complexi Security Analysis
overhead per Overhead per ty by its
block of data block of data strength
DES * 49392 for a56 bit Equal Exponenti chosen text only
key al
Model 1 ** 486 for a 72 bit Equal Exponenti Cipher text only
key al
Model 2 1953 for a 72 bit Equal Exponenti Cipher text ,known
key al plain text only
* Block of data refers to 64 bits.

** Block of data refers to 72 bits (9 character)
Table 6.5.2: Comparative study of the proposed Stream ciphers Algorithm 1: Model
3(Probabilistic encryption algorithm), Algorithm 2 with standard algorithm like
RC4 in terms of computational overhead, Data overhead, complexity and security
analysis.
Algorith Computational Data Overhead Complexit Security Analysis
m overhead per per 20 y by its
character character Text strength
RC4 * 64,000 for a 20 Exponential Chosen text only
40 bit key.
Model 3 2106 for a 72 bit 40 Exponential Chosen text only
key.
Algorith 500 for a 8 bit 20 Exponential Chosen text only
m2 key
* Reference from Applied Cryptography by Bruce Schiener

50000 80
70 Data overhead
40000
60 for a 20
30000 character text
50
Computational 40 character
20000 40
over head per text
block of data 30
10000 20 80 character
0 10 text
Model model DES 0
1 2 Model 1 Model 2 DES
Figure 6. 2: Pictorial Reresentation of the proposed encryption algorithms like

Model 1 & Model 2 with standard algorithm like DES in terms of computational
overhead & Data overhead.
70000 160
60000 140 Data overhead
50000 120 for a 20
100 character text
40000
Computational 80 40 character
30000 overhead per text
60
20000 character 40
10000 20 80 character
0 0 text
Model Model RC4 Model 3 RC4
3 4
Figure 6. 3: Pictorial Reresentation of the proposed encryption algorithms like

Model 3 & Algorithm 2 with standard algorithm like RC4 in terms of
computational overhead & Data overhead.
Thus in this work a comparative study of three developed models of algorithm 1
& a model of algorithm 2 is discussed in terms of computing power, complexity,
avalanche effect & security analysis on encryption process. In the given work, the onus is
on developing the subkeys, the encryption and decryption processes are inverse to one
another. Thus the computing power, Complexity, avalanche effect and security analysis
are equally valid for both encryption and decryption processes. And also Model 1&2 are
compared with block cipher like DES algorithm and Models 3 of algorithm 1&
Algorithm2 with standard stream cipher like RC4. The comparative study is represented
both in the form of tables and graphs.
Chapter 7
APPLICATIONS
7. Applications
7.1 Mobile Adhoc Networks:
In applications like Mobile Adhoc Networks [39], Wireless Sensor Networks and
Broad Casting Applications, New symmetric encryption techniques and probabilistic
encryption techniques can well be used. It can also be used for Key encapsulating
mechanism (KEM) along with Public key cryptography, which can well be used in secure
electronic transactions.
A mobile ad-hoc network (MANET) is type of wireless ad-hoc network used for
mobile routers. The routers are freely movable and can change from place to place. Such
type of systems can be used independently or may be connected to internet. They can be
used at different places and at different times, usually with all nodes connected to few
hops of each other, and having data transfer at a constant rate. A new substitution
algorithms, or in particular a probabilistic encryption can well be used in MANET
environment to provide security to the constant data sent among different nodes of each
hop. The Unique characteristics of MANET’s pose a number of non trivial challenges to
security design
1. Peer to Peer N/W architecture.
2. Shared wireless medium

3. No clear line of defense
4. Stringent Resource Constraints
5. Highly dynamic network topology.
In this environment, MANET architecture requires offloading of

heavy computation to more powerful devices. One more limitation is its vulnerability of
routing protocols that can freeze the whole N/W. One more requirement with MANET is
its combining of Key predistribution with secret sharing to achieve key distribution in
MANETs. Thus due to the absence of clear line of defense, a complete security solution
for MANET should integrate both proactive and reactive approaches.
7.2 A wireless sensor network
WSN is a wireless network [19,41,42] consists of distributed devices which

contains sensors to observe environmental conditions, such as temperature, sound,
vibration, pressure, motion or pollutants, at different places. The main application of this
sensor networks lies with motivated by military applications. Each node is associated
with sensors, with a radio transceiver, a small microcontroller, and a battery. The
capacity of the sensor considered depends on energy requirements and bandwidth and
computational speeds. A sensor network is supported with multi-hop routing algorithm.
Since the very important application of wireless sensor network lies with military
applications, security is of high importance.
Here are some constraints with security on Sensors
1. Slow computation rate.
2. Battery power tradeoff between security and battery life.
3. Limited Memory
4. High latency, consume power, turn on periodically.
Thus in this environments of MANET’s & WSN’s, there is a high possibility of pattern
matching to get any unauthorised information from the data being sent. To provide
security in this environment, the developed models can well be used. In probabilistic
model, multiple cipher texts will be generated for one plain text and any one cipher text
can be used for transmission. By the construction of the model, it is free from differential
and linear cryptoanalysis and also it is free from chosen cipher text attacks. In the second
algorithm, the encrypted data not only depends on the key but also on random nonce
value and a highly dynamic time stamp. Thus the probabilistic model and the time
stamped model avoids any statistical attacks and pattern matching attacks on the
encrypted data being sent, which makes the developed models suitable in these
environments. And more over the developed models require less computing power for
encryption & decryption process, they are well suited in these environments where the
data transfer is high.
7.3 Broad Casting Applications:
One more important application of encryption algorithms lies with Broadcasted

data [6]. Broadcasting applications like audio and video is transmitted to the public.
Television and radio programs are examples of broadcasting applications. In certain
applications like military applications, policing applications, financial applications
relevant to a larger variety of customers, message transmission through broad casting is
very much relevant. These are certain applications where higher level of security is
needed. Since the data rate is very high in broad casting applications, the computing
power needed for encryption process and decryption process is also high.
In this broadcasting environment the new symmetric encryption algorithms can

well be used. Model 1 with its low computing power and equal data overhead can well be
used in this environment. To provide better security, even Model 2 with its low
computing power, equal data overhead and variable length keys, can well be used. And
also the second algorithm with its small key length, variable nonce value and dynamic
time stamp can be used in broadcasting applications. The computing power of both the
algorithms is almost 80 percent less than the computing power required for standard
algorithms like DES & RC4 , the developed algorithms are suited in this environment.
The main disadvantage with this scheme is all the participants of this scheme have to
share the same private key for decryption purpose which purports the very concept of
security. This limitation can be overcome by having a key distribution centre where all
the participants have to be registered. The key distribution centre, based on the request
from the sender generates a session key which is distributed to all the participants of the
broadcasting mechanism. Using the session key and encryption algorithms, data transfer
is well secured in a broadcasting environment.
7.4 Key Encapsulating Mechanism:
Some more practical applications use hybrid encryption to deal with large
plaintext messages since the efficiency of the public key encryption algorithm is low and
computational head is high. As a main part of hybrid encryption schemes, Key
Encapsulation Mechanism [5,16] allows a sender to generate a random session key and
distribute it to recipient. In some communication scenario, the session key used in new
symmetric encryption model and as well probabilistic encryption models can be
encapsulated to designated group through their public keys. Anyone can encapsulate a
session key for a designated group and any recipient in the designated group can
decapsulate the session key with his private key. Thus the mechanism is secure against
adaptively chosen cipher text attacks in standard model, which provides ultimate security,
confidentiality and authentication to the transmitted data.
Chapter 8
SUMMARY & CONCLUSION
8.1 Summary
This study represents the importance of Encryption of data for storage and
transmission. The significance of encrypted data can be identified in light of the
mushrooming applications and globalization of communication. The advantages of
encrypting data manifest themselves in the form of security & confidentiality in real time
applications. Encryption of data is of particular significance in applications like email, e
commerce, e-cash where highly vulnerable communication lines is accessed for
transmission of highly volatile data.
The study traces the development of various encryption models in a real time
environment in all their breath taking diversity and breakthroughs in Chapters 2. The
significance of the advances and adaptabilities is measured in terms of their diversity of
applications in myriad ways that we feel in our daily lives.
The chapter 3 identifies the methodology used in the developed work. It is

classified as two algorithms. The first algorithm generates a sequence, followed by model
to generate sub keys and mapping of sequence or the sub keys on plain text to generate
cipher text. The second algorithm considers a key, a time stamp & a nonce value to
generate sub keys which are mapped on plain text to generate cipher text.
Chapter 3.2.1 builds a new block cipher algorithm. The algorithm considers a
matrix key and executes a sequence of steps which generates a sequence. A block of plain
text which is of equal length to the length of sequence generated is considered. Each
Character in the plain text is replaced by a corresponding numerical value added by a
value from the generated sequence .Thus the cipher text obtained becomes difficult to
break with out knowing the key. The strength of the generated model is studied in terms
of computing power, Avalanche effect, and complexity of the model in terms of
construction & strength. A Crypto analysis of the model is also identified to study the
security of the developed model.
Chapter 3.2.2 builds a new variable length key block cipher algorithm. The
algorithm considers a matrix key and executes a sequence of steps which generates a
sequence. Based on the similarity of values this sequence is being divided into basins.
The basins with minimum values will be eliminated. Remaining each basin represents
one block of data. Depending on starting input plain text character, corresponding basin
is considered as a key. The procedure is repeated for certain plain text depending on
chosen value. Thus the cipher text obtained becomes very difficult to be broken without
knowing the key. The strength of the generated model is studied in terms of computing
power, Avalanche effect, and complexity of the model in terms of its construction &
strength. A Crypto analysis of the model is also identified to study the security of the
developed model.
Chapter 3.2.3 builds a new stream probabilistic cipher algorithm. The algorithm
considers a matrix key and executes a sequence of steps which generates a sequence.
Based on the similarity of values this sequence is being divided into basins. Each basin
represents one character. Each character in the plain text is replaced by a set of basins
based on chosen base value. Each basin value is replaced by random values from basins
to generate the cipher text. Thus multiple cipher texts will be developed for one plain
text. Any one cipher text can be used for data transmission. Thus the cipher text obtained
becomes computationally infeasible to be broken with out knowing the key. The strength
of the generated model is studied in terms of computing power, Avalanche effect, and
complexity of the model in terms of its construction & strength. A Crypto analysis of the
model is also identified to study the security of the developed model.
Topic 3.3 builds a model which considers not only the key, but also a time stamp
and a nonce value to generate sub keys which are mapped to plain text to generate cipher
text. The strength of the generated model is studied in terms of computing power,
Avalanche effect, and complexity of the model in terms of its construction & strength. A
Crypto analysis of the model is also identified to study the security of the developed
model.
Trainings on the developed algorithms are discussed in chapter 4. Training an

encryption model involves variable length key, a plain text and an algorithm which
applies the key on the plain text to generate cipher text. This allows for the data to be
transmitted over the network in some form which cannot be read by any intruder. In the
first algorithm, the model is trained for different keys and an analysis of the models is
being done. The random matrix keys of the form 3*3 keys and 4*4 matrix keys are
considered in the training process. In the second algorithm, the model is trained for
different key values, time stamps & nonce values. By having small variations in the keys,
the models are studied for their increase in performance and working. The models are
also trained for their increase in security by having slight variations in the key values.
The models are also studied for changes in data overhead for small variations in the key
values.
A model to identify the garbled keys in transmission is discussed in chapter 5.
This facilitates for identifying and rectifying garbled keys while transmission through the
unreliable network.
A comparative study of all the developed models is discussed in chapter 6. All the
developed models are studied in terms of computing power, Avalanche effect and
complexity of the models in terms of their construction and strength. A crypto analytical
study of all developed models is also considered. Block Ciphers like Model 1 & Model2
of the first algorithm are compared with standard block cipher DES and ciphers like
Model 3 of the first algorithm & second algorithm are compared with stream cipher like
RC4.
Data encryption applications to communication networks are discussed in

Chapter 7. The applications covered are Mobile adhoc networks, wire_less sensor
networks and broad casting applications. The chapter highlights the significance of data
encryption to communication networks in light of the ever growing user demand for both
security and confidentiality in transmission of data/information. This is particularly
important in the face of computer access to financial transactions due to the impact of
Internet.
8.2 Conclusions.
Encryption models presented in chapters 3.2.1, 3.2.2, 3.2.3 &3.3 ranged over new
block cipher techniques to stream cipher techniques. The first algorithm considers a
matrix key and executes a sequence of steps which generates a sequence. This sequence
is used to generate four different encryption models.
Model 1 is a new symmetric encryption technique. In the Model it is observed

that, for a given key, the total computational overhead of the model1 is 486 calculations
The complexity of the model by its construction is O(n), by its strength it is exponential
in nature. It is also observed that by slight variations in the key a lot of variations in
cipher text is identified which provides more strength to the generated model. : This
algorithm is completely free from cipher text only, type of attack. By the other attacks,
the key may not be retrieved but a part of plain text may be retrieved.
Model 2 is a variable length key block cipher technique. It is observed that, for a
given key, the total computational overhead of the model2 is 1953 calculations The
complexity of the model by its construction is O(n2), by its strength it is exponential in
nature. It is also observed that by slight variations in the key and the plaintext a lot of
variations in cipher text is identified which provides more strength to the generated
model. This algorithm is completely free from cipher text only and known plain text
attacks but may not be completely free from chosen plain text and cipher text attacks
where a part of information may be leaked out by crypto analysis.
Model 3 is a stream probabilistic cipher algorithm. In this model, multiple cipher

texts will be generated for one plain text. Thus identifying the relationship between plain
text and cipher text is computationally infeasible. It is observed that, for a given key, the
total computational overhead of the model3 is 2025+81k calculations (where k refers to
number of computations for random pick up of values). The complexity of the model by
its construction is O(n2), by its strength it is exponential in nature. It is also observed that
by slight variations in the key and the plaintext a lot of variations in cipher text is
identified which provides more strength to the generated model. This algorithm is
completely free from cipher text, known plain text and chosen plain text and cipher text
attacks. This model is also free from public key type attacks.
Algorithm 2: It is a new stream cipher algorithm. It is observed that, for a given

key, time stamp &Nonce values, the total computational overhead is 500 calculations The
complexity of the model by its construction is O(nt), by its strength it is exponential in
nature. It is also observed that by slight variations in the key, time stamp & Nonce values
a lot of variations in cipher text is identified which provides more strength to the
generated model. This algorithm is completely free from cipher text, known plain text,
chosen plain text attacks.
In chapter 4 the developed algorithms are trained for different combinations of

keys. From the above study, it has been observed that the first algorithm is trained for a
[3*3] and [4*4} keys on a input ternary vector of 27 & 81 values with a base of 3, and on
100 values with a base of 4. For the input vectors of 27, 81 & 100 values, the number of
basins formed varies between 3 to 6. It has been observed that the more the values of the
vector, it does not guarantee in the increase of basins. But increase in the values of basins
are observed. It is also observed that for certain variations in the key, the model is
generating more basins. Thus for a certain random generated key, the number of basins is
evaluated. A better metric is provided to a random key which provides more number of
basins. Since the number of basins forms the characters of the alphabet, the more the
number of basins, the more will be the characters of the alphabet. A combination of
basins will increase the number of characters of the alphabet. It is also observed that by
an increase in the number of basins, the data overhead will be drastically reduced. The
analysis and results of the training model is illustrated in tables: 4.1 to 4.3. The second
algorithm is trained for different keys, time stamps and nonce values. It has been
identified that a lot of variations are observed by varying the above said values which
gives sufficient strength to the algorithm. In both the algorithms, the strength lies with
non uniformity of values in the sequence generated. The sequence generated by both the
algorithms is tested for various statistical tests like uniformity testing, Universal testing
and repetition tests. The results proved that the generated sequences by both the
algorithms withstand the above said statistical tests. The results are shown in fig. 4.1 &
4.2.
A model to identify the garbled keys in transmission is discussed in chapter 5.
This mechanism uses the same algorithm for the generation of sub keys and also for the
generation of sequence which is used to generate hash value. This hash value will be
appended to key while transmitted from the Key Distribution Centre. If any errors takes
place in the key while transmission, there will be a mismatch between received hash value
& generated hash value and accordingly the participating parties ask for retransmission of
key. This facilitates for identifying and rectifying garbled keys while transmission through
the unreliable network.
A comparative study of all the developed models is presented in chapter 6. All the
models are compared in terms of computing power, avalanche effect, complexity of
models in terms of their construction and strength and a crypto analysis of the models is
also undertaken. The summary of the study is presented in table: 6.4. The models are
compared with standard algorithms like DES & RC4 algorithms and the results are
shown in Tables: 6.5.1 & 6.5.2.
The applications of the proposed models are specified in chapter 7. Thus in

applications like Mobile Adhoc Networks, Wireless Sensor Networks and Broad Casting
Applications, New Symmetric encryption models and in particular probabilistic
encryption mechanism could well be used. The proposed models can also be used for
Key encapsulating mechanism (KEM) along with Public key cryptography, which can
well be applicable in secure electronic transactions and digital cash.
The conclusions drawn above, reflecting the overall security and confidentiality
rates of transmitting data, confirm the improvement in the efficiency of transmitting data.
The methodology used in the dissertation can be used for evaluating new encryption
algorithms in terms of multiple parameters. Further, the quantitative data indicates
relationship between Random key considered, sub keys (Basins) generated,
computational power needed by the first algorithm and the strength and security of the
algorithm. It also identifies the importance of multiple parameters like keys, time stamps
and nonce values used in second algorithm in terms of its security & strength.
In the case of Substitution encryption algorithms, the gain by using DES

algorithm is its low computational power, which will be very much gained by using the
developed models. The developed models are giving almost equal security at low
computational overhead (Computing power). As the security of encrypting models is
directly related to the key length [3], the more the key length the more will be the
security of the algorithm. But this parameter increases the computational overhead
(computing power) of the encryption algorithms. The security of the developed models
is relatively free from the key length which gives more flexibility regarding computing
power.
Another conclusion from the above study is freeness from public key attacks.
With probabilistic encryption algorithm (Model 3), a crypto analyst can no longer encrypt
random plain texts looking for correct cipher text. Since multiple cipher texts will be
developed for one plain text, even if he decrypts the message to plain text, he does not
know how far he had guessed the message correctly. Under this scheme, different cipher
texts will be formed for one plain text. Also the cipher text will always be larger than
plain text.
Thus the study discusses the distinguishing features of two encryption

algorithms and their models in terms of their applications. The first algorithm considers a
random matrix key which is multiplied with a ternary vector. A sign function is applied
on the product to generate a sequence. This sequence is used to build three different
encryption models. Each model can be used for encryption of data. Two models like
Model1 & Model2 are based on block cipher technique and Model3 is a stream cipher. In
Model 1, the sequence generated is used as subkey to be mapped to alphanumerical
equivalent of Plaintext to generate cipher text. The computational power of developed
model is low and the model is using sign function which makes it relatively free from
Differential cryptoanalysis. In model 2, the sequence generated is divided to basins based
on equal values. These basins contain unequal values and based on chosen value
corresponding basin is used as subkey. Since theses basins can be considered as variable
length keys, the model is relatively free from Linear cryptoanalysis also. In particular,
Model3 is based on probabilistic encryption scheme where multiple cipher texts will be
generated for one plain text and any one cipher text can be used for transmission of data.
Since the model is probabilistic in nature, it is free not only from Differential and Linear
cryptoanalysis but also free from Chosen Ciphertext attack. The model is also free from
public key attacks. The second algorithm considers not only the key but also time stamps
and Initialization Vector to generate a sequence which is used as subkey to generate
cipher text. The study outlines the components of these algorithms and their strength in a
real time environment.
This process of improvement has been spurred on by the ever-growing user

demand for security to the transmission (and storage) of information. The Postwar II
period ushered in a new multifaceted era of advances in science and technology;
emergence of multinational corporations; and globalization of economic activity. This
change has and will continue to put heavy user demand on security to data in
communication systems and networks. Increased competition globally in the
e_commerce and other applications underlines the need for secured communication. For
data storage and transmission, encryption becomes imperative for security and
confidentiality.
Chapter 9
FUTURE WORK
8. FUTURE WORK:
The present work deals with plain text being represented by numerical and
characters of English alphabet. The work can be improved so that it can support the
characters of not only English but also of other languages as well. The work can also be
improved to support not only text but also other forms of message transmission like
audio, video and images.
REFERENCES
1. Amjay Kumar, Ajay Kumar: Development of New Cryptographic Construct using

Palmprint Based Fuzzyvoult, EURASIP Journal on Adv. In Signal Processing, Vol 21, pp
234-238, 2009
2. Baocang Wang, Qianhong Wu, Yupu Hu: A Knapsack Based Probabilistic Encryption
Scheme, On Line March 2007, www.citeseer.ist.psu.edu.
3. Bluekrypt 2009: Cryptographic Key length Recommendations,

http://www.keylength.com
4. Blum L., Blum M , Shub M. : A simple unpredictable pseudo random number

generator , SIAM J. compute , 1986, 15, (2), pp 364-383.
5. Brics: Universally comparable notions of key exchange and secure channels, Lecture
Notes in Computer Science, Springer, Berlin, March 2004.
6. Sage.math.Washington.edu/home/jetchev/Public.html/docs/jetchev-talk.ppt- Broadcast
encryption schemes.
7. Brassard G.: Modern Cryptology , a tutorial lecture Notes on computer science , (325)
,(spring-verlas) .
8.Bruce Scheneier: Applied cryptography (John Wiley & sons (ASIA) Pvt. Ltd.
9. Carlone Fontaine & Fabien Galand: A Survey of Homomorphic Encryption for non
specialists, EURASIP Journal,Vol 07, Article 10.
10. Donavan G.Govan, Nathen Lewis: Using Trust for Key Distribution & Route
Selection in Wireless Sensor Networks, International Conference on Network Operations
& Management, IEEE Symposium 2008, PP 787-790.
11. Dorothy E. Denning et al.: Time Stamps in Key Distribution Protocol,

Communication of ACM, Vol 24, Issue 8, Aug 1981, pp 533-536.
12. E.C.Park, I.F.Blake: Reducing communication overhead of Key Distribution Schemes

for Wireless Sensor Networks: Computer Communications & Networks, ICCCN 2007,
pp 1345-1350.
13. Georg J.Fuchsbauer: An Introduction to Probabilistic Encryption, ‘Osjecki

Matematicki List 6(2006), pp37-44.
14. Guo D, Cheng L.M., Cheng L.L: A New Symmetric Probabilistic Encryption Scheme
Based on Chaotic Attractors of Neural Networks, Applied Intelligence, Vol 10, No.1, Jan
99, pp 71-84.
15. Hamid Mirvazri, Kashmiran Jumari Mahamod Ismail, Zurina Mohd. Hanapi:
Message based Random Variable Length Key Encryption Algorithm, Journal of
Computer Science, pp 573-578, 2009.
16. Hianyi Hu, Gufen Znu, Guanning Xu: Secret Scheme for Online Banking based on
Secret key Encryption, Second International Workshop on Knowledge Discovery & Data
Mining, Jan 23-25 2009.
17. Henry Baker and Fred Piper: Cipher systems(North wood books, London 1982).
18.J.William stalling :Cryptography and network security (Pearson

Education,ASIA1998)
19. Kaiping Xue: Study of improved key Distribution Mechanisms based on two length
structure for wireless sensor networks, International conference on adv. Information
Technology, 2008.
20. Krishna A.V.N.: A new algorithm in network security, International Conference Proc.
Of CISTM-05, 24-26 July 2005, Gurgoan, India.
21. Krishna A.V.N., Vishnu Vardhan.B.: Utility and Analysis of some Encryption
algorithms in E learning environment, International Convention Proc. Of CALIBER
2006, 02-04 Feb. 2006, Gulbarga, India.
22. Krishna A.V.N., S.N.N.Pandit: A new Algorithm in Network Security for data
transmission, Acharya Nagarjuna International Journal of Mathematics and Information
Technology, Vol: 1, No. 2, 2004 pp97-108
23. Krishna A.V.N, S.N.N.Pandit, A.Vinaya Babu: A generalized scheme for data
encryption technique using a randomized matrix key, Journal of Discrete Mathematical
Sciences & Cryptography, Vol 10, No. 1, Feb 2007, pp73-81
24. Krishna A.V.N., A.Vinaya Babu: Web and Network Communication security
Algorithms, Journal on Software Engineering, Vol 1,No.1, July 06, pp12-14
25. Krishna A.V.N, A.Vinaya Babu: Pipeline Data Compression & Encryption
Techniques in e-learning environment, Journal of Theoretical and Applied Information
Technology, Vol 3, No.1, Jan 2007, pp37-43.
26. Krishna A.V.N, A.Vinaya Babu: A Modified Hill Cipher Algorithm for Encryption
of Data in Data Transmission, Georgian Electronic Scientific Journal: Computer Science
and Telecommunications 2007 !N0. 3(14) pp 78-83.
27.Lester S. Hill, Cryptography in an Algebraic Alphabet, The American Mathematical

Monthly 36, June-July 1929, pp306–312.
28.Lester S. Hill, Concerning Certain Linear Transformation Apparatus of Cryptography,

The American Mathematical Monthly 38, 1931, pp135–154.
29. Maybec.J.S. (1981), Sign Solvability, Proceedings of first symposium on computer

assisted analysis and model simplification, Academic Press, NY.
30. M.Steiner, M.Waidner: Tutorial on Secure Electronic Commerce, 1999.
31. Pandit S.N.N (1963): Some quantitative combinatorial search problems. (Ph.D.
Thesis).
32.Pandit S.N.N (1961): A New matrix Calculus, J Soc., Ind. And Appl. Math. Pp
632-637.
33. Pci Yihting: A Temporal order Resolution algorithm in the multi server time stamp
service frame work, International Conference on Advanced Information Networking &
Applications, AINA 2005, Vol 2m 28-30 March, pp 445-448.
34. Phillip Rogaway : Nonce Based Symmetric Encryption,

www.cs.ucdavis.edu/rogeway.
35. R.S.Thore & D.B.Talange: Security of internet to pager E-mail messages (Internet for
India-1997IEEE Hyderabad section) pp.89-94.
36. Raja Ramanna Numerical methods 78-85(1990).
37. R.H.Rahnan, N,Nowsheen: A New Symmetric Key Distribution Protocol Using

Centralized Approach, Asian Journal of Information Technology, 2007 6(8) pp 911-915.
38. Suhas V. Patenkar Numerical Heat Transfer and Fluid Flow 11-75(1991).
39 Terry Ritter: Substitution Cipher with Pseudo-Random Shuffling, Cryptologia archive,

Volume XIV, issue 4, Oct 99, www.portal.acm.org
40. wikipedia.org/wiki/Mobile-adhoc-network
41. wikipedia.org/wiki/sensor_network
42. Wireless sensor networks: Edited by C.S.Raghavendra, Krishna M. Sivalingam, Taieb

Znati.

Great Encryption Algo PDF

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Great Encryption Algo PDF

Загружено:

Авторское право:

Доступные форматы

PERFORMANCE EVALUATION OF NEW

ENCRYPTION ALGORITHMS WITH

COMPUTER SCIENCE AND ENGINEERING

Under the Guidance of

Prof. A.VINAYA BABU

Department of Computer Science and Engineering

I hereby declare that the entire thesis work entitled, “PERFORMANCE

Date: 01 -08 - 2009

Department of Computer Science and Engineering

Acharya Nagarjuna University, Nagarjuna Nagar – 522 510

Andhra Pradesh, INDIA.

JNTUH, HYDERABAD, INDIA.

This is to certify that the present dissertation work entitled “P

Date: 01- 08 - 2009

HYDERABAD Prof. A.VINAYA BABU

I owe an inestimable debt of gratitude to Professor Dr. A.Vinaya Babu, Director,

2.1 Symmetric encryption schemes …

3.1 Problem Definition …

stamp for encryption mechanism.

3.3.1 Mathematical modeling of the problem …

4 Training of the Developed Model 86-103

7.1 Mobile Adhoc Networks …

8.2 Conclusion 130

9.1 Future work …

Publications of Author based on

Table No. 3.2.1.1: Encryption & Decryption process by Model 1…………………...44

Table No.3.2.1.2: Representation of No. of computations, Computing power,

Table No.3.2.2.1 Encryption & Decryption Process of Model 2.……………………54

Table No.3.2.2.2 Representation of No. of computations, Computing power,

Table No. 3.2.3.1 Encryption & Decryption Process of Model 3…………………...67

Table No.3.2.3.2 Representing No. of computations, computing power, complexity

Table3.3.2 Complexity by construction & strength of Model

Table 4. 6: Conversion of plain text to cipher text for a 10 character Text…………97

Table 6.1: Comparative Study of Number of computations and computational

Table 6.2: Comparative study of Different Generated Models in terms of Security

Table 6.3: Comparative Study of Different Generated Models by Avalanche effect,

Table 6.4: Summary of Comparative study of some models of the proposed

Figure 1.1: Encryption Model ………....…………………………………………........2

Figure 1.2: Symmetric Encryption Model…………………………………………….3

Figure 1.3: Public Key Encryption Model…………………………………………….4

Figure 1.4: Probabilistic Encryption Model ……….......……………….………........6

Fig 4.2: Sequential Data Representation of Algorithm 2……………...………… …103

Figure 6. 1: Pictorial Representation of Comparative study of some models of the

Figure 6.2: Pictorial Representation of the proposed encryption algorithms like

Figure 6.3: Pictorial Representation of the proposed encryption algorithms like

Plain text Cipher text

Figure 1.1 Encryption Model

Plain text Cipher text

Figure 1.2 Symmetric-key encryption

Encryption and decryption with a symmetric algorithm are denoted by

Plain text Cipher text

Figure 1.3 Public-key encryption

In public key encryption there is always a possibility of some information being

Plain text Multiple

Figure 1.4 Probabilistic encryption

An algorithm is unconditionally secure if, it is difficult to recover the plain text in

The complexity of an attack can be measured as Data Complexity, the amount of

As a thumb rule, the complexity of an attack is taken to be minimum of these

2.1 Symmetric Encryption Schemes:

2.1.1 Block ciphers

2.1.1.1 DES Algorithm:

2.1.1.2 Triple DES:

2.1.1.3 Homomorphic DES: