Revisiting AES Related-Key Differential Attacks with Constraint Programming

David Gérault (1) , Pascal Lafourcade (1) , Marine Minier (2, 3) , Christine Solnon (4)

1 LIMOS - Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes
2 LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
3 CARAMBA - Cryptology, arithmetic : algebraic methods for better algorithms
4 M2DisCo - Geometry Processing and Constrained Optimization

The Advanced Encryption Standard (AES) is one of the most studied symmetric encryption schemes. During the last years, several attacks have been discovered in different adversarial models. In this paper, we focus on related-key differential attacks, where the adversary may introduce differences in plaintext pairs and also in keys. We show that Constraint Programming (CP) can be used to model these attacks, and that it allows us to efficiently find all optimal related-key differential characteristics for AES-128, AES-192 and AES-256. In particular, we improve the best related-key differential for the whole AES-256 and give the best related-key differential on 10 rounds of AES-192, which is the differential trail with the longest path. Those results allow us to improve existing related-key distinguishers, basic related-key attacks and q-multicollisions on AES-256.

Mots clés: AES; Constraint Programming; Related-key attacks

Domaines: Intelligence artificielle [cs.AI]; Cryptographie et sécurité [cs.CR]

David Gérault, Pascal Lafourcade, Marine Minier, Christine Solnon. Revisiting AES Related-Key Differential Attacks with Constraint Programming. Information Processing Letters, 2018, 139, pp.24-29. ⟨10.1016/j.ipl.2018.07.001⟩. ⟨hal-01827727⟩