The Effect of Peer Selection and Buffering Strategies on the Performance of Peer-to-Peer File Sharing Systems

  • Authors:
  • L. Zou;E. W. Zegura;M. H. Ammar

  • Affiliations:
  • -;-;-

  • Venue:
  • MASCOTS '02 Proceedings of the 10th IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunications Systems
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Peer-to-peer systems contain a collection of equivalententities that form an application-layer overlay network.Each entity acts as client, server, and overlaynetwork router. Peer-to-peer systems enable the rapiddevelopment of distributed applications; in short order,peer-to-peer file sharing has become one of the mostpopular Internet applications. There are two primaryfunctions in a peer-to-peer file sharing application: contentdiscovery (determining which peers have the desiredcontent) and content retrieval (transferring contentfrom a peer to the requester). We focus on two aspectsof content retrieval that have received limited attention,namely peer selection and request buffering. Using acombination of analytic modelling and simulation, wefind that a simple greedy scheme that uses only local informationto select a peer has performance comparableto alternative schemes, including those that make moreglobally-informed decisions. We also find that the optimalrequest buffering strategy depends on the systemload and availability of peer information. For example,when peer information is scarce, download bufferingworks well; when peer information is plentiful, nobuffering is preferable for better throughput. Our resultshave implications for the design of high-performance,low overhead peer-to-peer file sharing systems.