Securing BitTorrent using a new reputation-based trust management system

Abstract

Nowadays, BitTorrent as a means of sharing files has become highly popular among internet users. However, due to the open nature of BitTorrent protocol and lack of any security mechanism, number of attacks against BitTorrent has significantly increased. Sybil, Collusion, Lying-Piece, Fake-Block, and Chatty-Peer are attack types which have been considered in this paper to secure BitTorrent against them. These attacks can decrease the download performance of BitTorrent clients considerably. In this paper a new reputation based trust management system to cover aforementioned attack types is presented. The proposed approach calculates a local score at peers and a global score at the tracker for each peer. First, peers are sorted according to their cumulative score at the tracker and then top 10 % of these peers are used to determine other peers global score. These local and global scores are used to find attackers. In addition, a novel formula has been utilized to calculate peers local score. Using the global score concept makes our mechanism robust and swift to detect collusion attack which has not been considered in most of similar previous works. In order to evaluate the effectiveness of the proposed system, several simulation and real experiments in the Emulab testbed were performed. The outcomes indicate that our method is highly effective in detection of rogue peers and Free-Riders; moreover, performance of honest peers has significantly improved.