Performance Modeling of BitTorrent Peer-to-Peer File Sharing Networks

Abstract: BitTorrent is undoubtedly the most popular P2P file sharing application on today's Internet. The widespread popularity of BitTorrent has attracted a great deal of attention from networking researchers who conducted various performance studies on it. This paper presents a comprehensive survey of analytical performance modeling techniques for BitTorrent networks. The performance models examined in this study include deterministic models, Markov chain models, fluid flow models, and queuing network models. These models evaluate the performance metrics of BitTorrent networks at different regimes with various realistic factors considered. Furthermore, a comparative analysis is conducted on those modeling techniques in the aspects of complexity, accuracy, extensibility, and scalability.