Improved Linear Sieving Techniques with Applications to Step-Reduced LED-64 - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Improved Linear Sieving Techniques with Applications to Step-Reduced LED-64

Résumé

In this paper, we present advanced meet-in-the-middle (MITM) attacks against the lightweight block cipher LED-64, improving the best known attacks on several step-reduced variants of the cipher in both single-key and related-key models. In particular, we present a known-plaintext attack on 2-step LED-64 with complexity of 248 and a related-key attack on 3-step LED-64 with complexity of 249. In both cases, the previously known attacks have complexity of 260, i.e., only 16 times faster than exhaustive key search. While our attacks are applied to the specific scheme of LED-64, they contain several general methodological contributions: First, we present the linear key sieve technique, which allows to exploit linear dependencies between key bits to obtain filtering conditions in MITM attacks on block ciphers. While similar ideas have been previously used in the domain of hash functions, this is the first time that such a technique is applied in block cipher cryptanalysis. As a second contribution, we demonstrate for the first time that a splice-and-cut attack (which so far seemed to be an inherently chosen-plaintext technique) can be used in the known-plaintext model, with data complexity which is significantly below the code-book size. Finally, we extend the differential MITM attack on AES-based designs, and apply it independently in two stages from both sides of the cipher, while using the linear key sieve and other enhancements.

Dates et versions

hal-01086175 , version 1 (22-11-2014)

Identifiants

Citer

Itai Dinur, Orr Dunkelman, Nathan Keller, Adi Shamir. Improved Linear Sieving Techniques with Applications to Step-Reduced LED-64. FSE 2014 - 21st International Workshop on Fast Software Encryption, Mar 2014, London, United Kingdom. pp.390-410, ⟨10.1007/978-3-662-46706-0_20⟩. ⟨hal-01086175⟩
174 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More