Security Evaluation of Lightweight Block Ciphers against Mixture Differential Cryptanalysis

The proliferation of the Internet of Things (IoT) has amplified the necessity for secure data transmission. Lightweight block ciphers are pivotal in fortifying the security of IoT systems, yet the resource-constrained nature of IoT often limits the complexity of their designs, especially the linear layer in typical Substitution-Permutation Network (SPN) and Feistel designs. This study investigates the vulnerability of seven lightweight block ciphers -specifically, CRAFT, Midori, SKINNY, MANTIS, LBlock, TWINE, and WARP, each catering to diverse application demands -concerning their susceptibility to mixture differential (MD) cryptanalysis. Modifying an automated tool based on linear programming, we identify MD distinguishers associated with these ciphers, exhibiting a higher number of rounds than that observed in the widely adopted AES block cipher. This disparity suggests that the simplified linear layer adopted by the lightweight block ciphers potentially compromises their resistance to MD distinguisher construction. Nevertheless, this compromise is counterbalanced by an augmented incorporation of rounds within the cipher design. Pertaining to the notion of security margins -denoting the percentage of rounds beyond the scope of constructed MD distinguishers relative to the total number of full rounds -the lightweight block ciphers subjected to scrutiny manifest elevated security margins, thereby demonstrating heightened resilience against MD cryptanalysis. This research provides a comprehensive security evaluation of the target lightweight block ciphers and introduces a versatile evaluation tool that can be adapted for the analysis of other aligned lightweight block ciphers in the context of MD cryptanalysis.