Data burst statistics and performance analysis of optical burst switching networks with self-similar traffic

The self-similar model can reflect the Internet traffic property more precisely than Poisson model, the theoretical and simulation results of burst length distribution based on the time threshold assembly mechanism are shown in this paper. Moreover, the simulation result of OBS schedule algorithm under self-similar traffic model is reported Both the analytic and simulation results show that burst length distribution is related to the length of assembly period; secondly, the assembly mechanism based on time period can reduce the self-similarity of traffic effectively; at last, the schedule algorithms in OBS have different performance based on the self-similar traffic.