计算机科学与探索2025，Vol.19Issue(10)：2844-2854,11.DOI:10.3778/j.issn.1673-9418.2409070
针对SKINNY算法的改进差分故障攻击方法
Improved Differential Fault Attack Method on SKINNY
摘要
Abstract
The existing differential fault attacks against the SKINNY algorithm have problems such as a large number of faults that need to be injected,high complexity in key searching,long attack execution time,and there may be remaining key in key space.To address the above issues,a new differential fault attack model based on double nibble/byte faults is proposed to implement differential fault attacks on the SKINNY-64-64 and SKINNY-128-128 algorithms.This model allows attackers to access and fully control the tweak(tweakable block cipher),assuming a fixed tweak for the tweak key input.By injecting double nibble and double byte faults into the R-3 and R-4 rounds of the key schedule algorithm for SKINNY-64-64 and SKINNY-128-128,respectively,the model recovers the main keys of SKINNY-64-64 and SKINNY-128-128 by establishing differential equations and computing equations according to constraints of the differential of S-box inputs in the last three rounds.The simulation results on software implementations show that,the proposed model can recover the main keys of SKINNY-64-64 and SKINNY-128-128 using only 5.9 and 6.7 pairs of faults,respectively.Compared with existing methods,the proposed model can maintain lower attack time overhead while injecting fewer faults to recover the unique keys of the algorithm.关键词
差分故障攻击/SKINNY/双半字节/字节故障/密钥编排Key words
differential fault attack/SKINNY/double nibble/byte fault/key schedule分类
信息技术与安全科学引用本文复制引用
武小年,杨济铭,张润莲..针对SKINNY算法的改进差分故障攻击方法[J].计算机科学与探索,2025,19(10):2844-2854,11.基金项目
广西重点研发计划(桂科AB23026131) （桂科AB23026131）
国家自然科学基金(62062026) （62062026）
桂林电子科技大学研究生教育创新计划项目(2024YCXS036).This work was supported by the Key Research and Development Program of Guangxi(Guike AB23026131),the National Natural Sci-ence Foundation of China(62062026),and the Innovation Project of Guilin University of Electronic Technology Graduate Education(2024YCXS036). （2024YCXS036）
