Cryptanalysis and design of symmetric primitives

Cryptographic primitives are the basic building blocks of various cryptographic systems and protocols. Their application is based on their well established properties. The security of a crypto system is proven under the assumption that the underlying cryptographic primitives provide some specific security levels. Therefore it is critical to use primitives that can meet these requirements. However, there is no general approach of constructing fast and provably secure primitives. Rather, the primitives undergo years of thorough cryptanalysis and only after no attacks have been found, they can be considered for real world applications.

This PhD thesis deals with the two main cryptographic primitives: block ciphers and cryptographic hash functions. The main contribution lies in presenting attacks on these algorithms. The analysis ranges from finding ad-hoc differential trails that are used for collision search and distinguishers on specific hash functions to automatic search tools that give the optimal differential trails for block ciphers. Weaknesses are shown for a number of SHA-3 candidates in the framework of rotational distinguishers and meet in-the-middle based preimage attacks.