Title :
An Effective Network-Aware Peer Selection Algorithm in BitTorrent
Author :
Qin, Fenglin ; Liu, Ju ; Zheng, Lina ; Ge, Liansheng
Author_Institution :
Sch. of Inf. Sci. & Eng., Shandong Univ., Jinan, China
Abstract :
This paper presents an effective network-aware peer selection algorithm to alleviate the topology mismatch issue in BitTorrent. By utilizing the characteristic of hierarchical network topology, the Tracker in BitTorrent classifies the peers into three clusters: local, intra-ISP and inter-ISP clusters based on peers´ traceroute results, and then chooses close peers within the first two clusters. Thus, the closeness among peers in BitTorrent can be enhanced. Experimental results show that, compared with standard BitTorrent, the proposed algorithm can achieve better download performance and greatly reduce both ISP´s backbone traffic and the cross-ISP traffic.
Keywords :
Internet; file organisation; peer-to-peer computing; BitTorrent; ISP; Tracker; hierarchical network topology; network-aware peer selection algorithm; Clustering algorithms; Delay; Information science; Intelligent networks; Internet; Network topology; Peer to peer computing; Signal processing algorithms; Spine; Telecommunication traffic; BitTorrent; Network-aware peer selection; P2P systems;
Conference_Titel :
Intelligent Information Hiding and Multimedia Signal Processing, 2009. IIH-MSP '09. Fifth International Conference on
Conference_Location :
Kyoto
Print_ISBN :
978-1-4244-4717-6
Electronic_ISBN :
978-0-7695-3762-7
DOI :
10.1109/IIH-MSP.2009.90