Parallel chaotic hash function based on the shuffle-exchange network

In recent years, cryptologists have been delving into chaos theory to design more secure cryptographic primitives. However, many existing chaos-based algorithms are slow due to floating point operations. They are mostly sequential in nature and therefore cannot take advantage of multicore processors for faster speed. In this paper, a new chaos-based hash function is proposed that utilizes multiple instances of chaotic maps that run in parallel to improve hashing speed. Parallelization is realized using the baseline network that also strengthens the security of the hash function due to its shuffling mechanism. The combination of linear and nonlinear chaotic maps is used to provide a high level of sensitivity to initial conditions, confusion and diffusion characteristics as well as strong collision resistance. Results show that the proposed design has strong security strength with near-perfect statistical qualities and fast hashing speed that surpasses both chaotic hash functions and the MD5 hash function.