Exploring and improving BitTorrent topologies

BitTorrent, the most popular peer-to-peer (P2P) file-sharing protocol, accounts for a significant fraction of the traffic of the Internet. Using a novel technique, we measure live BitTorrent swarms on the Internet and confirm the conjecture that overlay networks formed by BitTorrent are not locality-aware, i.e., they include many unnecessary long distance connections. Attempts to improve the locality have failed because they require a modification of the existing protocol, or interventions by Internet service providers (ISPs). In contrast, we propose a lightweight method that improves the locality of active swarms by 6% by suggesting geographically close peers with the Peer Exchange Protocol (PEX), without any modifications to the current system. An improvement of locality not only benefits the ISPs by reducing network transit cost, it also reduces the traffic over long-distance connections, which delays the need to expand the infrastructure, easing the power consumption. We expect that if used on a large scale our method reduces the Internet's energy consumption by 8 TWh a year.

[1]  Bryan Ford,et al.  Peer-to-Peer Communication Across Network Address Translators , 2005, USENIX Annual Technical Conference, General Track.

[2]  Xue Liu,et al.  Improving BitTorrent Traffic Performance by Exploiting Geographic Locality , 2008, IEEE GLOBECOM 2008 - 2008 IEEE Global Telecommunications Conference.

[3]  Farnam Jahanian,et al.  Internet inter-domain traffic , 2010, SIGCOMM '10.

[4]  B. Cohen,et al.  Incentives Build Robustness in Bit-Torrent , 2003 .

[5]  Pablo Rodriguez,et al.  Should internet service providers fear peer-assisted content distribution? , 2005, IMC '05.

[6]  Walid Dabbous,et al.  Pushing BitTorrent locality to the limit , 2008, Comput. Networks.

[7]  Stefan Schmid,et al.  Free Riding in BitTorrent is Cheap , 2006, HotNets.

[8]  Barath Raghavan,et al.  The energy and emergy of the internet , 2011, HotNets-X.

[9]  B. Huffaker,et al.  Distance Metrics in the Internet , 2002, Anais do 2002 International Telecommunications Symposium.

[10]  Fabián E. Bustamante,et al.  On blind mice and the elephant: understanding the network impact of a large distributed system , 2011, SIGCOMM.

[11]  Rodrigo Rodrigues,et al.  Adaptive Search Radius - Lowering Internet P2P File-Sharing Traffic through Self-Restraint , 2007, Sixth IEEE International Symposium on Network Computing and Applications (NCA 2007).

[12]  William Chan,et al.  Improving Traffic Locality in BitTorrent via Biased Neighbor Selection , 2006, 26th IEEE International Conference on Distributed Computing Systems (ICDCS'06).

[13]  Songqing Chen,et al.  TopBT: A Topology-Aware and Infrastructure-Independent BitTorrent Client , 2010, 2010 Proceedings IEEE INFOCOM.

[14]  Fabián E. Bustamante,et al.  Taming the torrent: a practical approach to reducing cross-isp traffic in peer-to-peer systems , 2008, SIGCOMM '08.

[15]  Stephen J. Wright,et al.  Power Awareness in Network Design and Routing , 2008, IEEE INFOCOM 2008 - The 27th Conference on Computer Communications.

[16]  Marcel Dischinger,et al.  Detecting bittorrent blocking , 2008, IMC '08.

[17]  Abraham Silberschatz,et al.  P4p: provider portal for applications , 2008, SIGCOMM '08.

[18]  Michael Sirivianos,et al.  Free-riding in BitTorrent Networks with the Large View Exploit , 2007, IPTPS.

[19]  Matteo Varvello,et al.  Traffic Localization for DHT-Based BitTorrent Networks , 2011, Networking.

[20]  David E. Culler,et al.  PlanetLab: an overlay testbed for broad-coverage services , 2003, CCRV.

[21]  David Mazières,et al.  Kademlia: A Peer-to-Peer Information System Based on the XOR Metric , 2002, IPTPS.