A Study on the Efficiency of File Sharing in Peer-to-Peer Networks

During the past few years, there has been an increasing interest in peer-to-peer networks. A wide variety of applications have been proposed to take advantage of high scalability and abundant resources at peer nodes. File sharing applications, such as Napster and Gnutella, are among the most popular ones that have received considerable attention from researchers. However, a number of problems need to be solved to make these applications work more efficiently. In this project, we will focus on two important and closely related problems: file availability and adaptive topology. The availability of a file is defined as the number of complete files that can be found among all peers in the network. Increasing the file availability can reduce the time required to download a file. In the first year, we will propose a file distribution protocol to improve the file availability through peer clustering, which has been studied extensively in the literature. Compared with existing approaches, the proposed protocol aims to distribute the blocks of a file in a timely and effective manner. In addition, since peer-to-peer networks are highly dynamic, the proposed protocol will deal with node dynamics. One of the major issues that may affect the efficiency of file sharing is peers’ voluntary actions. In the second year, we will design an adaptive topology protocol which can be used by peers to maintain connections with good peers. Moreover, the proposed protocol has intrinsic incentive for peers to share resources. As a result, the resulting topologies are believed to be more efficient than existing protocols. We will evaluate the performance of the proposed file distribution protocol and the adaptive topology protocol over various types of network topologies through simulations.

Project ID:PB9907-12667
External Project ID:NSC99-2221-E182-053