一种基于完全性的不可能差分区分器构造方法

摘要

基于混合运算的密码算法(MOC)以安全性高、软硬件实现效率高等特点受到人们的广泛关注.完全性指输出的每一比特都包含有输入每一比特的信息,达到完全性是密码算法设计的一个基本原则.该文提出针对MOC算法完全性分析的通用算法,并在此基础上提出利用完全性寻找MOC算法的不可能差分区分器的方法,此构造方法可直接给出MOC算法高重量的不可能差分区分器且搜索效率高,为MOC算法不可能差分区分器的实际构造提供了理论指导和技术支持.应用此方法找到了 SIMON 系列算法全部现有的最长不可能差分区分器,并找到了SPECK系列算法更多的不可能差分区分器.%Mixed Operation based Ciphers (MOC) attract cryptographers owing to their high security and high efficiency on both software and hardware platforms. As a basic principle of cryptosystem design, completeness refers to that every output bit contains the information of every input bit. This paper presents a universal algorithm of completeness analysis against MOC. Based on the algorithm, a method of constructing impossible differential distinguishers utilizing completeness is proposed. This method constructs heavy weight impossible differential distinguishers directly with high efficiency. The method can provide theory and technology direction for the construction of impossible differential distinguishers. Then, this paper analysis SIMON and SPECK with this method and introduces all the longest impossible differential distinguishers of SIMON currently public and new impossible differential distinguishers of SPECK.