Design of a Predictor for MD5 Based Cryptographic Systems: A TVAC-PSO Based Approach

This paper puts forward an efficient and optimized method to determine the probable-colluders in a MD5 based cryptosystem. Along with the conventional MD5 implementing architecture, our scheme employs a predictor control block which takes the message stream from the user, and provides the log-list of the equal length bit-streams that are most likely to produce collisions with the message stream in a real time system. The predictor controller uses Time varying Acceleration Coefficient Particle Swarm optimization (TVACPSO) algorithm. Besides proposing the algorithm itself, the paper also surveys the performance of the predictor when employed with different hardware platforms.