Modeling and performance analysis of BitTorrent-like peer-to-peer networks
Proceedings of the 2004 conference on Applications, technologies, architectures, and protocols for computer communications
Improving Traffic Locality in BitTorrent via Biased Neighbor Selection
ICDCS '06 Proceedings of the 26th IEEE International Conference on Distributed Computing Systems
Should internet service providers fear peer-assisted content distribution?
IMC '05 Proceedings of the 5th ACM SIGCOMM conference on Internet Measurement
P4p: provider portal for applications
Proceedings of the ACM SIGCOMM 2008 conference on Data communication
TopBT: a topology-aware and infrastructure-independent bittorrent client
INFOCOM'10 Proceedings of the 29th conference on Information communications
Understanding peer distribution in the global internet
IEEE Network: The Magazine of Global Internetworking
| Hi-index | 0.00 |
Peer-to-peer (P2P) applications, most notably BitTorrent (BT), are generating unprecedented traffic pressure to the Internet Service Providers (ISPs). To mitigate the costly inter-ISP traffic, P2P locality, which explores and promotes local peer connections, has been widely suggested. Unfortunately, existing proposals generally require that an ISP control the neighbor selection of most peers, which is often not practical in a real-world deployment given there are noncooperative trackers. In this paper, for the first time, we examine the characteristics and the impacts of these noncooperative trackers through real-world measurements. We find that tracker blocking has the potential to address this noncooperation problem, and help the ISPs to control more peers for traffic locality. Yet, how to guarantee torrents' availability at the same time remains a significant challenge for the ISPs. To this end, we model the tracker blocking problem coherently with torrent's availability, and address it through a novel selective tracker blocking algorithm, which iteratively improves traffic locality with a given availability threshold. Our trace-driven evaluation shows that our solution successfully reduces the cross-ISP traffic in the presence of noncooperative trackers and yet with minimal impact to torrents' availability.