《Invited lecture》

  • Date : 2017.12.5(Tue)
  • Jiageng Chen (Central China Normal University)


Improved Differential Characteristic Searching Methods


The success probability of differential and linear cryptanalysis against block ciphers heavily depend on finding differential or linear paths with high statistical bias compared with uniform random distribution. For large number of rounds, it is not a trivial task to find such differential or linear paths. Matsui first investigated this problem and proposed a solution based on a branch and bound algorithm in 1994. Since then, finding good concrete differential or linear path receives less attention than it deserves. In this paper, we revisit the differential attack against several S-Box based block ciphers by carefully studying their differential characteristics. Inspired by Matsui's algorithm, we provide an improved solution with the aid of several searching strategies which enable us to find the best differential characteristics for the three investigated ciphers (LBlock, TWINE, and PRESENT) efficiently. Furthermore, we provide another way to evaluate the security of ciphers against differential attacks by comparing the strength of the ciphers from differential characteristic's point of view. We also investigate the accuracy when using the number of active S-Boxes to evaluate security margins against differential attacks, which is the common method adapted when new ciphers are designed. We further apply the methodology to the KATAN family and derive the best attack so far.