Research On P2P Content Distribution Algorithm

According to the statistic collected in recent years, peer-to-peer (P2P), which is the most predominant application of internet family, carries over 60% internet traffic, and the main part of which is consumpted by content distribution application. So P2P content distribution is a focus issue.As the specific applications of P2P content distribution, P2P file distribution and P2P media distribution are studied in this paper, and their typical algorithms'analysis and improvement are conducted.P2P file distribution, which carry the largest part of P2P traffic, has always been on hot spots. BitTorrent system, as the typical representative of P2P file distribution, uses tit-for-tat (TFT) algorithm as peer selection algorithm (chocking algorithm) to guarantee download speed and fairness, which means decisions as to which peers to unchoke are based strictly on current download rate. But this choking algorithm results in slow starts for downloading and deficient uploading links when peers have downloaded almost all of blocks. Therefore, an adaptive choking algorithm is proposed, in which peers create additional uploading links when there are a few uploading links. Experimental results show that, the adaptive algorithm can effectively accelerate the downloading launch rate, increase utilization of uploading links, so as to enhance file distribution performance of BitTorrent system.P2P media streaming distribution, with nearly media streaming network expansion, gradually aroused the concerns of industry and academia. P2P media streaming solutions based on data-driven is widely used for its simple implement, efficience, robust and flexibility. This paper studies DONet, the typical representative of data-driven overlay network; points out that under highly dynamic networks, scheduling algorithm of DONet can't guarantee continuity for lack of rapid resources location; presents a scheduling algorithm termed SA-SC (Scheduling Algorithm Based on Stage Classification), block at different time stage is applied to different scheduling algorithm, so as to rapid location for blocks at urgert stage and full collaboration for blocks at other stages; simulation results show that, SA-SC is a guaranteed QoS scheduling algorithm, even under highly dynamic networks, SA-SC can still guarantee the continuity of streaming media and collaboration between peers.