construction of side channel attack resistant s-boxes using genetic algorithms based on coordinate functions

Background and objectives: substitution-box (s-box) is one of the essential components creating confusion and nonlinear properties in cryptography. to strengthen a cipher against various attacks, including side channel attacks, these boxes need to have numerous security properties. in this paper, a novel s-box construction method is introduced aimed at improving the resistance of s-boxes against power analysis attacks. methods: in the preprocessing phase of this approach, a suitable initial s-box with some basic security properties was generated by adopting a fast algorithm. then, in the main stage, using the initial s-box, we generate new s-boxes which not only have the properties of the initial s-box but also have significantly improved under another set of security properties. to do this, new s-boxes were generated using a genetic algorithm on a particular subset of the linear combination set of coordinate functions of the initial s-box. results: the performed experiments demonstrated that the values of all security properties of these new s-boxes, especially the measures of transparency order, signal-to-noise ratio, confusion coefficient, bijection property, fixed point, and opposite fixed points, have been substantially improved. for example, our experiments indicate that 70, 220, 2071, 43, and 406 s-boxes are found better than the initial s-box, respectively, in the dimensions of 4×4 through 8×8. conclusion: in this paper, a new s-box construction method is introduced where the properties related to side channel attacks are improved, without destroying other security features. besides, some results obtained from generated s-boxes in the dimensions of 4×4 through 8×8 demonstrated that the generated s-boxes are not only improved relative to the initial s-box, but also in certain cases, considerably better than some well-known s-boxes.