故障模型下MORUS算法的差分扩散性质研究

Research on Differential Diffusion Property of MORUS in Fault Model

MORUS算法是由H.Wu等人设计的一类认证加密算法,目前已顺利进入CAESAR竞赛第3轮竞选.研究MORUS算法故障模型下的差分扩散性质.采用面向比特的随机故障模型,结合差分分析技术与中间相遇思想,改进了针对MORUS算法的差分链搜索算法.运用该算法找到了5步概率为2^(-85)的差分链,从而实现了对初始化过程5步的简化版MORUS-640-128算法的差分-区分攻击,攻击所需的数据量和区分优势分别为2^(89)和0.99965.最后,利用差分故障分析方法对认证过程3步的简化版MORUS-640-128算法进行了伪造攻击.

MORUS is a third-round CAESAR candidate of authenticated cipher designed by H. Wu et al. With a fault model,the diffusion property of MORUS is analyzed in this paper. By using a bit-oriented random fault model,the search algorithm for the differential chain of MORUS is improved with the usage of differential analysis and meet-in-the-middle technique. Through this algorithm,a 5-step differential chain is discovered with a probability of 2^-85. The differential-distinguish attack on the initialization of 5-step reduced version of MORUS-640-128 is proposed with the data complexity of 2^89 and the distinguishing advantage of 0.99965. By using differential fault analysis method,the forgery attack on 3-step authentication of MORUS-640-128 is formed.