Stochastic fluid model for P2P content distribution networks

Abstract

Traditional content distributions such as Web cache and content distribution networks need to reconfigure the networks or purchase additional servers to support these services. While P2P content distribution networks do not change the underlying infrastructure. We propose to build the multi-P2P networks based on the current unstructured and structured P2P networks. And the contents are distributed between multi-P2P networks. We explore the stochastic fluid analytical models for the purpose of multi-P2P networks performance. And we get the results of the relation between the content get probability and the number of multi-P2P networks, number of nodes, users' request rate, content expiration rate through this analytical model. The number of multi-P2P networks has a more impact on the content get probability than that of the universal P2P network. It has been proved that the scheme of constructing multi-P2P networks is effective.