CONSTRUCTION OF STRONG S-BOXES FOR BLOCK CIPHERS

Abstract

This thesis deals with the construction of cryptographic primitives for symmetric cryptosystem. The symmetric cryptosystem consists of stream cipher and block cipher. The Cryptographic primitives used as a fundamental building block in the stream cipher and block cipher are Boolean functions and vectorial Boolean functions (S-Boxes). In this thesis, our primary goal is to design Boolean functions and S-Boxes with strong cryptographic properties. The majority portion of our research work is devoted to construct cryptographically strong S-Boxes for block cipher using evolutionary techniques and chaotic maps. We develop a new tool called BSAT (Boolean function and S-Box Analysis Tool) for evaluating the cryptographic strength of the Boolean functions and S-Boxes. This tool serve as a standalone tool and plays an important role when the Boolean functions and S-Boxes are evolved using evolutionary computing techniques. We implement some important cryptographic properties in an optimized manner to reduce computational complexity. We design an improved hybrid Genetic algorithm by integrating local search procedure into GA to construct balanced Boolean function to optimize the nonlinearity and autocorrelation value. The produced Boolean function attains the nonlinearity and autocorrelation value as a pair are (26, 16), (116, 24), (488, 40), and (2002, 72), for input size 6, 8, 10, and 12 variable, respectively. We apply a new cost function and analyze the di↵erent fitness/cost functions to observe how the fitness/cost function choice improves the algorithm for getting the optimal results.