Abstract :
[en] We present the first automatic search algorithms for the best related-key differential characteristics in DES-like ciphers. We show that instead of brute-forcing the space of all possible differences in the master key and the plaintext, it is computationally more efficient to try only a reduced set of input-output differences of three consecutive S-box layers. Based on this observation, we propose two search algorithms – the first explores Matsui’s approach, while the second is divide-and-conquer technique. Using our algorithms, we find the probabilities (or the upper bounds on the probabilities) of the best related-key characteristics in DES, DESL, and s^2DES.
Scopus citations®
without self-citations
16