CWI Cryptology Group

Abstract: Collision attacks on SHA-1 are for a major part based on combinations of local collisions. At EUROCRYPT'13 we introduced a new exact and holistic method to analyze combinations of local collisions that takes dependencies between them fully into account. This talk will discuss current work in extending this work to the remaining part of collision attacks towards improved SHA-1 collision attacks. We also discuss how this method can be applied in the counter-cryptanalysis of SHA-1 to significantly speed up the collision detection algorithm for SHA-1.

Abstract: The cyberwarfare malware Flame was, among other things, notable for its ability to spread inside local networks via Windows Update, disguised as a security update from Microsoft. Using a chosen-prefix collision attack, based on differential cryptanalysis, on the cryptographic hash function MD5, the attackers obtained a certificate in Microsoft's PKI which they used to sign the update. Even though the attackers created a new variant collision-attack, it has some similarities to MD5 collision attacks that can be found in the literature. We attempt to reconstruct the collision attack and the underlying differential path families and illustrate our reconstruction techniques on example.
Joint work with Marc Stevens

Abstract: In this talk we shall discuss the meet in the middle technique that was introduced by Merkle & Hellman in their seminal paper that showed that Double-DES is not more secure than single DES. Their technique was later used by Chaum & Evertse to attack reduced-round variants of DES. While being left out for several years, the technique is recently enjoying a renaissance, and become an active research field, with application to many block ciphers.
Joint work with Itai Dinur, Nathan Keller, and Adi Shamir.

Abstract: In this talk we shall discuss some of the new results on meet in the middle attacks, and mainly the recent work by Dinur et al. We will cover the new and surprising results in the field of the meet in the middle attacks, that can be applied to a much larger range of problems than multiple encryption.
Joint work with Itai Dinur, Nathan Keller, and Adi Shamir.