ATTI DELLA “FONDAZIONE GIORGIO RONCHI” ANNO LXVI, 2011 - N.

COMPLEXITY

Determination of complexity of pseudo-random

sequence based on Z-transform
SATTAR BADER SADKHAN (*),
RAGHAD KADHIM SALIH (**)

SUMMARY. – In recent years, stream cipher systems have played a big role especially in computer
networks. Stream cipher systems depend on pseudo-random (PR) binary key sequences which are
mixed with the plaintexts using addition with modulo two to produce the ciphertexts. The PR
key sequences are characterized by three properties which define the measure of security for these
sequences. These properties are period, complexity and randomness. There are several methods to
determine the degree of complexity of these PR sequences like Berlekamp-Massey method, matrices
techniques, and using software computing techniques (Neural Networks, Fuzzy Logic and Genetic
Algorithms). This paper presents a proposed method to determine the degree of complexity of these
PR sequences using one of the properties of the Z-transform which is periodic sequence property.
This proposed method enable us to compute the degree of complexity of any periodic sequence
produced from linear or nonlinear generators and accurate results are obtained. The steps of the
proposed procedure are given, two different examples are illustrated, one of them for a PR sequence
over GF(2), and the second example for PR sequence over GF(5).

Key words: Stream cipher, Pseudo random sequence, complexity evaluation, Z-transform.

1. Introduction

Cryptography and information security are considered as one of important

sciences in the world, especially after using the wide spread use of computers
in many world wide applications. The need to keep certain messages secret has
been appreciated for thousands of years. The idea of a cipher system is to disguise
confidential information in such a way that its meaning is unintelligible to an un-
authorized person. The information to be concealed is called plaintext, while the
resulted information is called ciphertext (1).

(*) University of Babylon, IRAQ; e-mail: [email protected]

(**) Department of Applied Sciences, University of Technology, Babylon, IRAQ
166 S.B. Sadkhan - R.K. Salih

One of the well known cipher systems is the stream cipher system. Stream
cipher system is very important system for the information security. In stream
cipher system the pseudo random (PR) sequence (i.e. the key) is combined or
mixed with the plaintext using modulo 2 addition to produce the ciphertext. It is
absolutely crucial that if the key of the cipher system is known, one can determine
the plaintext from the ciphertext. Hence the key or the PR sequence must have
long period, high complexity and randomness properties to have acceptable secu-
rity. The complexity of the PR sequence is the length of the minimum linear feed-
back shift register (LFSR) that can generate the sequence. The LFSR of length n
can be characterized by the characteristic polynomial f(x):
f (x) = c 0 + c1x + c 2 x 2 + ...+ c n−1x n−1 + x n
where c0, c1, …, cn–1 are 0 or 1. The PR sequence must have high complexity since
for a sequence with a known complexity L, we need 2L consecutive bits to deduce
the entire sequence, since if 2L consecutive bits are given, one can write a system
of L-equations in L unknown variables and find its unique solution (1,2).
James L. Massey (3) suggested an algorithm which is known as Berlekamp-
Massey algorithm for computing the complexity of the PR sequences and J.M.
Baker and P. Hughes gave a new explanation of the Berlekamp-Massey algorithm
using a method based on matrices (2).
The aim of this paper is to provide a proposed method for complexity evalu-
ation of the Pseudo-Random Sequence using the Z-transform. The steps of the
proposed procedure are given, two different examples are illustrated one of them
for a PR sequence over GF(2), and the second example for PR sequence over
GF(5).

2. The Z-transform

The Z-transform is an important tool for analyzing linear discrete time sys-
tem. It is used to transfer sequences of numbers into algebraic equations of the
variable z which, in many cases, helps the solution of problems (4).
The Z-transform of a sequence of numbers {f(k)} which is identically zero for
negative discrete times (i.e., f(k) = 0 for k = –1, –2, –3,…) is defined by
∞
[1] Z { f (k)} = F(z) = ∑ f (k)z −k
k=0
where z is an arbitrary complex number (4,5).
The Z-transform possesses many notable properties. Table 1 lists the salient
properties enjoyed by the Z-transform.
 Determination of complexity of pseudo-random … 167

Table 1
Properties of Z-Transform

Property Discrete Sequence Z-Transform

1 Linearity af (k) + bg (k) aF (z) + bG (z)
2 Right-Shifting f (k – m) z–mF (z)
zN
3 Periodic-Sequence f (k) = f (k + N) F(z) = F1(z)
zN −1
k
4 Convolution ∑ f1(k − i) f 2 (i) F1(z) F2 (z)
i=0

m−1
5 Left-Shifting f (k + m) z m F(z) − ∑ f (i)z m−i
i=0

The Z-transform opens up new ways of solving the problems. It is well

known that Z-transform is used successfully in many engineering problems. Some
applications of Z-transform are (5):

1. Solution of the linear difference equation.

2. Transfer function.
3. Digital Filter Design.
4. Frequency Response.

3. Proposed method for complexity evaluation

Periodic sequence property of the Z-transform is used to determine the com-

plexity of the PR sequence (which is used as the key sequence in stream cipher
system).
A sequence of numbers which repeats itself every (N) discrete-time units is
said to be periodic with period N. Such sequences satisfy the property:

[2] f (k) = f (k + N)

for all nonnegative k.

The Z-transform of the first period of the periodic sequence characterized
by relationship [2] is
N−1
[3] F1(z) = ∑ f (k)z −k
k=0
168 S.B. Sadkhan - R.K. Salih

Since this first period is repeated every N discrete time units, it follows by the
right shifting property that the Z-transform of the periodic sequence is given by:
F(z) = F1(z) + z −N F1(z) + z −2N F1(z) + z −3N F1(z) + ...

(
= F1(z) 1+ z −N + z −2N + z −3N + ... )
where z–mNF1(z) designates the Z-transform of the mth period of the periodic se-
quence. The infinite sum within brackets is readily shown to be given by:
∞ ∞ m zN
∑ z −mN = ∑ (z −N ) for z > 1
=
m=0 m=0 zN −1
Hence, the Z-transform of the periodic sequence becomes
zN
F(z) = F1(z) for z > 1
zN −1
This property enables us to find the complexity L of the periodic PR se-
quences in stream cipher system as follows (6):

1. Evaluate p1 (z) and p2 (z) where:

p1 (z) = zN F1 (z)

p2 (z) = zN – 1.

2. Compute P(z), Q(z) and F(z) as follows:

p1(z)
P(z) =
gcd [ p2 (z), p1(z)]
p2 (z)
Q(z) =
gcd [ p2 (z), p1(z)]
where gcd [p2(z),p1(z)] is the greatest common divisor of two polynomials p2(z)
and p1(z), and
P(z)
F(z) = .
Q(z)

3. Evaluate the complexity L of the PR sequence by:

L = deq [Q(z)],

where deq [Q(z)] is the degree of the characteristic polynomial Q(z).

Example 1
Consider the following Pseudo Random sequence Si over GF(2):
 Determination of complexity of pseudo-random … 169

Si = 0010111,

where i = 0, 1, …, 6 and the period N = 7.

In this example the Z-transform is proposed to determine the complexity of
the sequence Si from the degree of the characteristic polynomial Q(z).
Hence, by taking the Z-transform to Si we get:
6
p1(z) = z 7 ∑ f (k)z −k = z 5 + z 3 + z 2 + z
k=0

p2 (z) = z 7 − 1
gcd [ p2 (z), p1(z)] = z 4 + z 2 + z + 1
and
z P(z)
F(z) = 3
= .
z + z +1 Q(z)
Then, the complexity L of the PR sequence Si is:
L = deq [Q(z)] = 3.

Example 2:

Consider the following sequence Si over GF (5):

Si = 2,1,4,3,1,0,0,1,2,4,1,1

i = 0,1, …,11, with period 12.

The Z-transform is used to compute the complexity of the sequence Si as
given in Appendix 1. Also the Z-transform F(z) is given in Appendix 1.
Then, the complexity L of the PR sequence Si is:
L = deq [Q(z)] = 4

4. Conclusions

Deciphering of the ciphertext in stream cipher system depends on the avail-

ability of the key of the ciphertext. So, one of the important properties of the
PR key sequence is to have high complexity in order to be difficult for the cryp-
tanalyst to obtain the entire sequence when only small segment of it is known.
There are several methods for computing the complexity of the PR sequences as
Berlekamp-Massey method. The Z-transform was used properly to compute the
complexity of the PR key sequences that determines the ability of security of these
sequences. The results of the Z-transform are accurate to determine the complex-
170 S.B. Sadkhan - R.K. Salih

ity for any real or binary PR sequence with any period. Moreover, the proposed
method is computed easily in digital computer.

Appendix 1
11
p1(z) = z 12 ∑ f (k)z −k = 2z 12 + z 11 + 4z 10 + 3z 9 + z 8 + z 5 + 2z 4 + 4z 3 + z 2 + z
k=0
p2 (z) = z 12 − 1 = z 12 + 4

gcd [ p2 (z), p1(z)] = z 8 + z 7 + 3z 6 + 3z 5 + 3z 4 + 4z 3 + 2z 2 + 2z + 1

2z 4 + 4z 3 + 4z 2 + z P(z)
F(z) = = .
z 4 + 4z 3 + 3z 2 + 2z + 4 Q(z)

REFERENCES

(1) BAKER H.J., PIPER F.C., Cipher Systems: The Protection of Communications (Northwood
Publications, London 1982).
(2) BAKER J.M., HUGHS P.M., Communications Speech and Vision, J. Proc. I, IPIDDG 136,
1989.
(3) MASSEY J.L., Shift Register Synthesis and BCH Decoding, IEEE Trans. on Information
Theory, IT-15 (1), January 1969.
(4) CADZOW J.A., Discrete-Time Systems, (Englewood Cliffs, N.J., Prentice Hall, Inc.,
1973).
(5) OGATA K., Modern Control Engineering, 3th Ed., (Printed in New Jersey, 1997).
(6) R.K. SALIH, Analysis of Pseudo Noise Generator Design for Communication Systems
using State Space Method, M.Sc. thesis, Univ. of Technology, Baghdad, June 2004.