Performance bounds analysis on different scheduling algorithms for networks with self-similar traffic

There are kinds of Internet applications such as VoIP, online gaming and video Conferencing which need quality of service guarantees making the performance analysis critical, and the traffic of such applications usually exhibits the properties of self-similarity and heavy tail. In this paper, we use the generalized Stochastically Bounded Burstiness model to characterize the network traffic with self-similar property and self-similar property. And we derive the delay bounds for such kind of traffic in the system under three different scheduling algorithms: First In First Out, Static Priority and Generalized Processor Sharing. We utilize the relationships between the traffic-amount-centric stochastic arrival curves and the virtual-backlog-centric stochastic arrival curve as well as the relationships between the weak stochastic service curve and the stochastic service curve making the analysis of the performance bounds of the through traffic in the tandem systems feasible.