A MATHEMATICAL MODEL FOR LONG RANGE DEPENDENT PEER TO PEER TRAFFIC AT THE INDIVIDUAL PEER

摘要

In this paper we propose a new mathematical model for P2P traffic at the individual peer. The model is based on Alternating Fractal Renewal Process (AFRP) such that the ON/OFF periods are similar to generalized Zipf distribution as we define. The download/seed rates are considered constant. Modeling traffic in P2P network is important, because recent measurements have shown that 60% of internet traffic is P2P traffic in 2004. Further more, better understanding of traffic helps us in QoS provisioning, bandwidth allocation and congestion control. Some models have been proposed for P2P network modeling. But their concentration is on performance issues and traffic measurement. Subhabrata Sen and Jia Wang have experimentally studied the Zipf distribution in P2P traffic characterization. Yunfei Zang has also shown empirically the long range dependent property in some P2P traffic. Different form their studies our model not only can be used to characterize P2P traffic but also captures Long Range Dependence (LRD). Simulation studies show the validity of our proposed model.