Downloader-initiated random linear network coding for peer-to-peer file sharing

Network coding has recently drawn tremendous attention. Applying network coding to Peer-to-Peer (P2P) file sharing is still a relatively new area of study. Adopting network coding in a P2P file sharing network can make data distribution more effective and relieve the burden of the original seeder. In this paper, we first provide the mathematical analysis modeled from contemporary P2P file sharing protocols to analyze the efficiency of network coding in P2P file sharing networks. Then, we further investigate the "unlucky combination" problem encountered by the traditional random linear network coding algorithm when applied to a P2P file sharing network. Hence, we propose a "Downloader-Initiated Random Linear Network Coding" algorithm, which lets the downloader determine the random linear coding to eliminate the "unlucky combination" problem, and therefore further improves the performance of P2P file sharing networks.