Non-Gaussian characteristic and FARIMA(p,d,q) traffic models

Extensive measurements of real-life traffic demonstrate that the probability density function of the traffic has a non-Gaussian feature. If a traffic model cannot capture this characteristic any analytical or simulation results will not be accurate. Hence, this paper studies the impact of non-Gaussian traffic on network performance, and presents an approach that can accurately model the marginal distribution of real-life traffic while accounting for both the longand short-range dependence. We validate our promising procedure by simulation.