A new dna based approach of generating keydependentmixcolumns

International Journal of Computer Networks & Communications (IJCNC) Vol.7, No.2, March 2015
DOI : 10.5121/ijcnc.2015.7208 93
A NEW DNA BASED APPROACH OF GENERATING KEY-
DEPENDENTMIXCOLUMNS TRANSFORMATION
Auday H. Al-Wattar1
, Ramlan Mahmod2
,Zuriati Ahmad Zukarnain3
, and NurIzura
Udzir4
,
1
Faculty of Computer Science and Information Technology, Universiti Putra Malaysia,
43400 UPM, Serdang, Selangor.
ABSTRACT
The use of key-dependent MixColumns can be regarded as one of the applied techniques for changing the
quality of a cryptographic algorithm. This article explains one approach for altering the MixColumns
transformation engaged in the AES algorithm. The approach employed methods inspired from DNA
processes and structure, which relied on the key.The parameters of the proposedMixCloumns have
characteristics identical to those of the original algorithm AES besides increasing its resistance against
attack.The original transformation uses single static MDS matrix while the proposed methods used
dynamic MDS. The security of the new MixColumns was analyzed, and the NIST Test Suite tests were used
to test the randomness for the block cipher that used the new transformation.
KEYWORDS
MixColumns, Block cipher, DNA, AES, NIST, MDS
1.INTRODUCTION
Recently, the need for anextremelyefficientapproach to attaininformation security is critical and
vital. Cryptography has been and stays behind the most proficient approach employed to achieve
security. Rijndael is a symmetric block cipher that was selected by (NIST) (National Institute of
Standards and Technology), [1] in 2001 as (advanced Encryption Standard, FIPS 197) AES.
In general, it is based on repeated rounds of transformation that alter input plaintext to
encryptedtext or ciphertext output. Each round consists of several functions and always includes a
depending on the roundsecretkey. Multiple roundsestablish inverse transforming ciphertext into
the original, using the same secret key. AES has 128 –bit block size, and a key length is 128, 192
or 256 bits, relying on to the number of rounds for the algorithm. It is using a byte array known
asthe state of (4x4)size in each cycle of the encryption / decryption process. The majority of the
algorithm calculations is accomplished in finite fields [2,3].
At AES, the MixColumns transformation is the most important function within the linear unit of
symmetric encryption algorithms.Itis themajor source of diffusion in the AES block cipher. Each
column is dealt with as a polynomial through GF (2 ), and then modulo + 1is multiplied by a
fixed polynomial c (x) =3 + + + 2.The inverse of this polynomial is ( ) = 11 +
13 + 9 + 14. The MixColumnsprocedure can be executed by

94
multiplying a coordinate vector of four numbers in Rijndael's Galois field by the
following circulated MDS matrix:
The multiplication operation is performed as complicated operation using the multiplied while the
addition operation is a simple XOR operation, as the math of the operation is madein Rijndael's
Galois field. Figure 1 shows the AES MixColumntransformation[4].
Figure 1: AES MixColumn transformation
The linear and differential cryptanalysis isthecriticall,demandingg for AES algorithm, where, it
can beissuedswitho standard techniques of differential and linear cryptanalysis. From the analysis
of resistance to differential and linear cryptanalysis, it was deduced that the random, unknown
and key-dependent permutation transformation is considered as a good feature in enhancing the
resistance of block cipher against the differential and linear attacks, since these attacks are need
known transformations. The detailed properties of substitution and permutation functions,
specifically the structure that is completely dynamic and unknown to the cryptanalyst, assist and
support the block cipher to be resistance to attacks. For the attacker, differential and linear trail
over numerous rounds is regarded as afundamental requirement, and in the reality of key-
dependent transformations output differences rely on additional key used. In addition, for any
additional different key values, there are unrelated differentials over multiple rounds and
therefore diverse linear differential trails. As a result, it is difficult for the attacker to utilize the
current linear and differential techniques of cryptanalysis.
Although many previous works on enhancing the security of the AES block cipher against
attacks, no singlework has proposed a MixColumns that is designed or created using DNA bio-
inspired techniques.
This paper proposed a new technique for obtaining a powerful key-dependent
MixColumnsdepending on operations that have been inspired from really biological DNA
processes. Subsequently, it tested the new transformationusing the NIST randomness tests for the

95
cryptosystem that with the DNA-based MixColumn and performed a security analysisof the
proposed transformation.
2. DNA BACKGROUND
DNA (Deoxyribo Nucleic Acid) is a molecule that characterizes the genetic material for every
living organism. It is assumed as the genetic drawing of living or existing creatures. A single-
strand of DNA consists of a chain of molecules known as bases, defined as four letters {A, C, G,
and T}[5,6]. One of the most fundamentalcharacteristics of the DNA strand series is that it is
oriented; as a result, ATCGTACT is distinct from TCATGCTA.
The DNA strands exist as pairs as (G) associated with (C) and (A) associated with (T)
figuringcomponents named base pairs. The reverse DNA strands representing the opposite of the
strand bases; for example, AGCTAGGCATAA becomes AATACGGATCGA, while the
complement of the strands can be represented as A¯ ≡ T and C¯ ≡ G. Therefore, GCATAA will
become TTATGC[7,9]. The DNA segment consists of two parts named as Exon, and Entron
respectively.
CENTRAL DOGMA
The central dogma process of the DNA strand bases is one of the main methods that characterize
the biology DNA system. It comprises of operations on DNA and RNA (Ribonucleic acid),
including transcription, translation, and replication. The DNA segment consists of parts:
Coded parts (Exon) and non-coding parts (Intron).
The process of the central dogma of molecular biology can be described in Figure 2,
in which the genes’ information glides into proteins [10].
FIGURE 2. DNA BASES REVERSE COMPLEMENT TECHNIQUE
 According to Figure 2, there are two methods within the central dogma as
transcription and translation which achieved to obtain the Protein. The
transcription methods are performed by removing the non-coding parts
(Introns) and keeping the coding part (Entrons), while the translation
methods including converting the RNA as MRNA into Proteins.
In this paper, the creation of a key-dependent MixColums is based on and inspired
by transcription process.

96
3. THE PROPOSED METHOD
In this paper the proposed MixColumns transformation will refer to as a key-dependent
MixColumns transformation ( _ ).
This ( _ ), willrely on the key cipher ( ), which is differentin each round (r).
In the original AES, the MDS matrix of the MixColumns transformation
remainspermanent for all rounds, whilethe proposed method the MDS matrix will be
changed for each round according to the key cipher.
To explain the proposed method, we first used the AES original MDS matrix as a base
matrixas shown in Figure 3.
Figure 3: A 4×4 MDS
There are two ways to use the key Kr for the ( _ ), The first one is to use four bytes of the
key such that have taken the values ranging between 0, and 3, while the second way is to use only
one byte of the key. In the both ways the key values are stored withinMOK matrix.
The applying of MOK on the MDS matrix to generate a new MDS matrix for each round is
inspired by the real DNA transcription process as shown in Figure 2. According to this procedure
the skipped columns will be regarded as Introns while the left behind columns will be considered
as Exon. The values of MOK will state the number of columns that will be skipped representing
the number of Intron. The generating method which called columns transcription is performed by
skipping a number of columns of the MDS matrix according to the amount exist in the MOK
First:MOK with four bytes (values):
For the first way (four bytes), the following procedure will be performed to obtain the ( _ ),
The work of this process will be done according to the key stored in MOK, and the values of the
MDS matrix‘s first row.
The columns of this matrixwill be shifted according to the key values as following:
If the key value = 1, then the row’s values of theorigin MDS matrix will be shifted one positionto
the right.
If the value of the key = 0, then the row’svalues of MDS will not change.

97
If the key value = 2, thenthe valuesof MDS rows willbe shifted to the right 2 positions.
Finally, if the value of the key = 3, then the MDSrow values will be shifted 3 positions.
Figure 4, demonstrates the whole process. Since the values of the key are changingat each
round, this will lead generating new different MDS for MixColumns transformation for
every round.
Figure 4: Generating new key-dependent MDS matrix (MOK) matrix
Second:MOK with one byte (value):
In this case thevalue within the MOK(which could be 0, 3) as one byte of the key will
specify a certain MDS matrix as showing in Figure 5,
Figure 5: Generating new key-dependent MDS matrix (MOK) matrix
The value of the key in MOK matrix indicates the number of columns of the first row of MDS
that should be skipped, referring to biology, DNA concept the skipped columns are considered as
(Intron) which are removed, while the remaining columns are considered as (Exon) through the
transcription/ splice process

98
4. TEST AND RESULT
4.1 Correlation coefficient
The correlation coefficient involves values ranging from (-1 and +1). According to[11] the
following values are in a good range for explicating the correlation coefficient as stated in Table
1.Note that this paper considers the two variables of plaintext (p) and ciphertext (c)
Table 1. Accepted range values for interpreting the correlation coefficient
Equation 1illustrates the applying of the correlation coefficient functions:
E(c) =
1
s
p (1)
where: s is the entire number of bits, p ,c are the chains of s measurements for p and c, p is bits
value of plaintext, c is bits value of ciphertext, E(c) is mathematic anticipation of c.
The variance of p can be expressed by equation 2,
D(p) =
1
s
[p

− E(p)] (2)
Lastly, the related coefficients r can be expressed by equation 3,
r =
E{[p − E(c)][c − E(c)]}
D(p) D(c)
(3)
The experiment examined theAES block cipher using the new ( _ ). The Scatter chart of
the results is presented in Figure 6. It illustrates that the majority of correlation values, at different
rounds through the whole block cipher are near to 0, which indicates a strong positive (or
negative) non-linear relationship. Only 0.007% are near to +1 or -1, in round 1, representing a
weak positive (or negative) non-linear relationship. From the results, it can be inferred that the
block cipher has an improved confusion performance.

99
Figure 6. Scatter chart of the correlation test results on whole block cipher
4.2 NIST Suite Randomness test
The randomness test is one of the security analysis to measure the confusion and the diffusion
properties of the new encryption algorithm, as carried out in [12-14][15] and[16].
NIST Suite [17] is a statistical test suite for randomness by NIST used to assess the cryptographic
algorithm. The suite test assesses whether the outputs of the algorithms under certain test
condition exhibit properties that would be expected randomly generated outputs.
In order to evaluate the randomness of ciphertext, an experiment that included a set of data as
random plaintext and random 16 byte key in the ECB mode was conducted. The 100 sequences of
1,059,584 bits were constructed and examined. The list of statistical tests applied throughout the
experiments is illustrated in Table 2.
Table 2. Breakdown of 15 statistical tests applied during experimentation

100
The proportion of sequences that passed a specific statistical test should lie above the proportion
value p, defined in Equation (4).
p∝ = (1−∝) − 3
∝ (1−∝)
n
= (4)
Where∝ is the significant value, n is the number of testing sequences.
For this experiment, the proportion value is:
p∝ = (1 − 0.01) − 3
0.01(1 − 0.01)
100
= 0.960150 (5)
Where n= 100, and ∝=0. 01.
The p-value readings for each round constructed is illustrated in Figure 7. This figure
demonstrates the randomness test for 15 statistical tests of block cipher that used the proposed
MixColumnsfor three rounds. From this figure, at the end of the second and third rounds, all of
the 41 statistical tests fall over 96.0150%, which is evidence that the output of the algorithm is
completely random.
Figure 7. Randomness tests results of whole block cipher for 3 rounds
4.3 CRYPTANALYSIS
The new MixColumns transformation is dynamic and its changing at each round according to
round key-value this feature make the job harder for the attackers since the analysis of dynamic
unit is more difficult than the static one. For the dynamic MixColumns, the attacker has the
possibility of 2n
!which consider a high number, that increasing the resistance of the algorithm

101
against the attacks.The AES security factors plus the new security features applied by the
proposed (KdM_Tr), are worked to enhance security of the block cipher.
5. CONCLUSION
In this paper a new dynamic key-dependent MixColumns transformation for AES block cipher
was proposed with chosen byte or four bytes of the private key. The new transformation is not
fixed, but changeable at each round according to the round key values. The MDS matrix of the
MixColumns is changed according to this key. This unit was tested with the correlation
coefficient and the 15 statistical randomness tests of NIST Test Suite. The analyzing of the
obtained results showed that, this new transformation is more secure and resistance against the
attacks, on which concluded that it is potential to employ it for an encryption. This will increases
the stability of AES against linear and differential cryptanalysis.
REFERENCES
[1] J. Daemen and V. Rijmen, "AES proposal: Rijndael," in First Advanced Encryption Standard (AES)
Conference, 1998.
[2] V. Rijmen and J. Daemen, "Advanced Encryption Standard," Proceedings of Federal Information
Processing Standards Publications, National Institute of Standards and Technology, pp. 19-22, 2001.
[3] L. Information Technology, Announcing the Advanced Encryption Standard (AES) [electronic
resource]. Gaithersburg, MD :: Computer Security Division, Information Technology Laboratory,
National Institute of Standards and Technology, 2001.
[4] P. FIPS, "197: Advanced encryption standard (AES)," National Institute of Standards and
Technology, 2001.
[5] M. Zhang, M. X. Cheng, and T.-J.Tarn, "A mathematical formulation of DNA computation,"
NanoBioscience, IEEE Transactions on, vol. 5, pp. 32-40, 2006.
[6] M. Zhang, C. L. Sabharwal, W. Tao, T.-J.Tarn, N. Xi, and G. Li, "Interactive DNA sequence and
structure design for DNA nanoapplications," NanoBioscience, IEEE Transactions on, vol. 3, pp. 286-
292, 2004.
[7] E. R. Kandel, "The molecular biology of memory storage: a dialogue between genes and synapses,"
Science, vol. 294, pp. 1030-1038, 2001.
[8] G. I. Bell and D. C. Torney, "Repetitive DNA sequences: some considerations for simple sequence
repeats," Computers & chemistry, vol. 17, pp. 185-190, 1993.
[9] A. G. D'yachkov, P. L. Erdös, A. J. Macula, V. V. Rykov, D. C. Torney, C.-S. Tung, et al., "Exordium
for DNA codes," Journal of Combinatorial Optimization, vol. 7, pp. 369-379, 2003.
[10] F. Crick, "Central dogma of molecular biology," Nature, vol. 227, pp. 561-563, 1970.
[11] K. Wong, "Interpretation of correlation coefficients," Hong Kong Medical Journal, vol. 16, p. 237,
2010.
[12] F. Sulak, A. Doğanaksoy, B. Ege, and O. Koçak, "Evaluation of randomness test results for short
sequences," in Sequences and Their Applications–SETA 2010, ed: Springer, 2010, pp. 309-319.
[13] Q. Zhou, X. Liao, K.-w.Wong, Y. Hu, and D. Xiao, "True random number generator based on mouse
movement and chaotic hash function," information sciences, vol. 179, pp. 3442-3450, 2009.
[14] V. Patidar, K. K. Sud, and N. K. Pareek, "A Pseudo Random Bit Generator Based on Chaotic Logistic
Map and its Statistical Testing," Informatica (03505596), vol. 33, 2009.
[15] J. Soto and L. Bassham, "Randomness testing of the advanced encryption standard finalist
candidates," DTIC Document2000.
[16] V. Katos, "A randomness test for block ciphers," Applied mathematics and computation, vol. 162, pp.
29-35, 2005.
[17] E. B. Smid, S. Leigh, M. Levenson, M. Vangel, A. DavidBanks, and S. JamesDray, "A Statistical Test
Suite for Random and Pseudorandom Number Generators for Cryptographic Applications."

102
Authors
AudayH.Al-Wattar obtained his B.Sc. degree in Computer Science from Mosul University and his M.Sc.
Degree from Technology University-Baghdad in 2005 .He presently pursing his Ph.D. From Universiti
Putra Malaysia, Malaysia under the guidance of Prof. Dr. Ramaln bin Mahmod at Computer Science and
Information Technology Faculty. He works as lecturer at Mosul University (since 2005), at Computer
Science and Mathematics Faculty - Computer Science Department. His area of interest includes Computer
Security, Programming languages.
Ramlan B Mahmodobtained his B.Sc. in computer Science, from Western Michigan, University, U.S.A. in
1983, and M.Sc. degree in Computer science, from Central Michigan, University, U.S.A. and Ph.D. degree
in Artificial Intelligence from, Bradford University, UK in 1994. His previous workings
Experience/Position are as:
System Analyst, PETRONAS,1979-1980. Lecturer,Mathematic Department, Faculty of Science, UPM,
1985-1994. Lecturer,Department of Multimedia, Faculty of Computer Science & Information Technology,
UPM, 1994 – 2002. Associate Professor,Department of Multimedia, Faculty of Computer Science &
Information Technology, UPM, 2002 – 2010.
Deputy Dean,Faculty of Computer Science & Information Technology, UPM, 1998 – Nov
2006.Dean,Faculty of Computer Science & Information Technology, UPM, 2010-2013. Professor, Faculty
of Computer Science & Information Technology, UPM, 2012 - now. His research interest includes Neural
Network, Artificial Intelligence, Computer Security and Image Processing.
NurIzurabintiUdzirobtained her B.Sc. in computer Science, from UniversitiPertanian Malaysia, in 1995,
and M.Sc. degree in Computer science, from Universiti Putra Malaysia, in 1998, and Ph.D. degree in
Computer Science from, University of York, UK in 2006.
Associate Professor,HeadDepartment of Computer Science, Faculty of Computer Science & Information
Technology, UPM. Her research interest includes Computer security, secure operating systems, Access
control, Distributed systems, Intrusion detection systems.
ZuriatiBinti Ahmad Zukarnain obtained her B.Sc. in Physics,, from Universiti Putra Malaysia, in 1997,
and M.Sc. degree in Information Technology, from Universiti Putra Malaysia, in 2000, and Ph.D. degree in
Quantum Computation and Quantum Information from, University of Bradford, UK, 2006,UK in 2006.
Associate Professor, Department Of Communication Technology and Networking, Faculty of Computer
Science & Information Technology, UPM. Her research interest includes information, computer and
communication technology (ICT), Quantum Information Systems and Distributed Systems, Quantum
Computing, Computer Networks and Distributed Computing.

A new dna based approach of generating keydependentmixcolumns

More Related Content

What's hot (18)

Viewers also liked (19)

Similar to A new dna based approach of generating keydependentmixcolumns (20)

More from IJCNCJournal (20)

Recently uploaded (20)

A new dna based approach of generating keydependentmixcolumns