自动化搜索ARX密码差分特征的方法

Automatic search algorithm for differential characteristics in ARX ciphers

要解决ARX密码算法差分特征的自动化搜索问题,关键是要解决搜索过程中模加差分的快速计算。首先,提出了相关差分分布表的概念,通过查找相关差分分布表,可以有效地计算模加的差分以及差分概率;其次,利用相关差分分布表,将Matsui算法扩展到ARX密码,提出了自动化搜索ARX密码差分特征的算法;最后,利用提出的搜索算法,搜索SPECK算法的差分特征,得到了SPECK32、SPECK48和SPECK64的最优差分特征。

How to compute the differential of modular addition efficiently is critical to automatic search for differential characteristic in ARX ciphers. To solve this problem, firstly, the concept of collerated difference distribution table（CDDT） was proposed. By looking up the CDDT, it was very efficient to compute the differential probability of modular addition. Secondly, extending Matsui＇s algorithm to ARX cihpers and using CDDT, an automatic search algorithm was proposed, and the algorithm could give the differential characteristic with highest probability in ARX ciphers. Finally, the proposed algorithm was applied to the ARX cipher SPECK, and got the best differential characteristics for SPECK32, SPECK48 and SPECK64.