Evolutionary Computing for Designing Cryptographic Primitives for Block Cipher: Challenges and Opportunities

Abstract

The cryptographic primitives such as S-Box and permutation layer are primary building blocks for designing block cipher. Since S-Box is the only nonlinear component in a block cipher, it is essential to select secure S-Boxes. The security of any block cipher depends upon the cryptographic properties of an S-Box and the lower bound of the number of active S-Boxes. Apart from S-Boxes, there are several other cryptographic primitives such as permutation layer plays a significant role in increasing the security of block cipher. Since the search space is huge for constructing \(n \times m\) S-Box, it is very difficult to find a good S-Box. The problem of finding and evolving an S-Box is an optimization problem. The purpose of our research work is to give a systematic review of how evolutionary techniques can be applied for constructing such cryptographic primitives, challenges of achieving optimal results, and opportunities for applying new techniques or fine-tuning several control parameters to improve the performance of an existing algorithm. In this paper, we discuss the genetic algorithm and memetic algorithm for constructing an \(n \times m\) bijective S-Box with important cryptographic properties. We address the challenges and future direction of the currently ongoing research.