Sukhinin

In this article we propose new generators of uniformly distributed pseudorandom sequences based on cellular automata. The advantages of such generators are good statistical properties of output sequences and high performance of hardware implementation. We justify the choice of the structure and parameters of generators and describe details of classical and non-uniform cellular automata usage in the structure of generators. We also compare hardware implementations of our generators and eSTREAM candidates.

We analyze the dependence of avalanche effect in classical and non-uniform cellular automata with balanced transition functions on the choice of cell neighborhood. We introduce space and integral characteristics to describe avalanche effect and also introduce a notion of optimal avalanche effect. We show that non-uniform cellular automata have better characteristics given the same neighborhood cardinality. The obtained results can be used in development of different algorithms such as block ciphers, hash functions and pseudorandom numbers generators.