Reducing Download Times in Peer-to-Peer File Sharing Systems with Stochastic Service Capacities

Abstract

The main problem for an individual user peer in a peer-to-peer network with heterogeneous source peers is the peer selection problem, namely, switching among source peers and finally settling on one, while keeping the total time of probing and downloading to a minimum. There has been little investigation on selecting source peers with stochastic service capacities. The main contribution of this paper is to address the problem of reducing download times in peer-to-peer file sharing systems with stochastic service capacities. A precise analysis of the expected download time is given when the service capacity of a source peer is a random variable. A chunk-based switching and peer selection algorithm using the method of probing high-capacity peers is proposed and the expected download time of the algorithm is analyzed. Two sub problems of the optimal choice of the threshold of high-capacity source peers and the optimal order of probing are also solved. The performance of the algorithm is compared with the random chunk-based switching method. It is shown that noticeable performance improvement can be obtained.