New Technique to Improve BitTorrent Performance Based on Application Layer Traffic Optimization

Abstract

Currently, BitTorrent is the most popular protocol that is widely used for P2P file sharing system. Nevertheless, BitTorrent protocol exposes some inefficient processes that are choosing neighbours and selecting pieces. These are mainly based on the knowledge of overlay topology without considering underlying internet topology that might cause of traffic problems such as the bottleneck at some point of network. Therefore, we propose the new approach for BitTorrent protocol with Tracker Localized Algorithm, Picker Localized Algorithm and Chocker Localized Algorithm which autonomous hops between peers are calculated to make efficient decisions. Then to compare the AS hop between 5 AS and 7 AS in the network. We conducted simulation of our scheme based on the PeerSim. The simulation results show that the peers provide better piece downloading on BitTorrentunderlying topology and interacting among peers more efficiently. The Tracker localized algorithm outperforms the original BitTorrent for reducing the average downloading time of 20%. The breaking time of peers is the least when using Chocker localized algorithm. Additionally, the results show that ours scheme can decrease the peer-to-peer traffic and optimize traffic distribution for the whole network.