Cryptographic hash functions are one of the most important tools used in the design of efficient security protocols and systems such as digital signatures that enable secure communications over the Internet.
Cryptographic protocols are often proved secure assuming that the underlying hash function is secure. The hash function SHA-1 was adopted as an American standard in the early 1990s, and it has been deployed in many cryptographic protocols since then.
New cryptanalytic approaches have since dramatically reduced the confidence in SHA-1 and other similar hash functions. Therefore the US Government (NIST) initiated a global competition in 2007 to find a new secure hash function which will be denoted SHA-3. The DTU cryptology group is a co-designer of the proposal Grøstl which has advanced to the third and final round of the competition together with four other systems.
This project is about analysis and design of cryptographic hash functions. The most important competitors of Grøstl share the same characteristics. They are all dedicated designs using only three elementary operations:
-
modular additions
-
exclusive-ors
-
rotations of words.
All those operations are extremely fast on modern processors and such designs outperform Grøstl when it comes to speed. However, very little is known about the security of such designs. There is no solid theory that would give any argument for the resistance against standard differential attacks. This is illustrated by a long history of broken designs of that kind. Results in this project will likely have implications also for encryption systems, like block ciphers and stream ciphers that use some of the same design principles.
Contact persons: Søren S. Thomsen and Lars R. Knudsen