Analysis And Design Of Hash Function

Hash function plays a significant role in the field of information security. The classic hash functions, MD4-x series, had been fully attacked by kinds of cryptanalysis during 2004, which gives rise to the rapid development of analysing and designing new hash functions.It is a common way to design components of hash function based on mixing operations of different groups. In this thesis, based on the characteristic of the input difference, we analyse the differential propagation property of the mixing operations, then we use the theory of rational power series on free monoid to study the linear approximation property of the mixing operations. Also some applications are given.For attacking hash functions, it mainly refers to attack the iterative structure and the delicated MD4-X series hash functions. Our thesis first discusses herding attack of hash functions, and then we propose a new attack technique based on the naive iterative structure and the backward prediction property of compression function. We call it inverse herding attack due to its original name herding attack and give some instances for attacking some hash algorithms. Meanwhile some weakness in choosing constants in delicated hash functions such as MD5 and SHA-1 has been pointed out and another kind of compression function proposed at FSE 2006 are also analysed especially its permutation component.Block-cipher based principle has been introduced in designing hash functions. We propose a new hash function algorithm-RDH based on the framework of block cipher Rijndael. Mixing operations of different groups and rotation technique guarantee sufficient confusion and diffusion between message and chain variables. The design principle and security analysis are also included in this thesis.