Multiobjective Evolution of Hash Functions for High Speed Networks

Hashing is a critical function in capturing andanalysis of network flows as its quality and execution timeinfluences the maximum throughput of network monitoringdevices. In this paper, we propose a multi-objective lineargenetic programming approach to evolve fast and high-qualityhash functions for common processors. The search algorithm simultaneously optimizes the quality of hashing and the executiontime. As it is very time consuming to obtain the real executiontime for a candidate solution on a particular processor, theexecution time is estimated in the fitness function. In order todemonstrate the superiority of the proposed approach, evolvedhash functions are compared with hash functions available in theliterature using real-world network data.