Inproceedings
An analysis of BitTorrent cross-swarm peer participation and geolocational distribution
Contribution Summary
This paper presents an analysis of BitTorrent cross-swarm peer participation and geolocational distribution. The authors collected a large dataset of 2 terabytes from 16 swarms of popular TV shows, including Breaking Bad, Dexter, and True Blood. The data was gathered over a month-long period, resulting in 1.27 billion peer hits and 6.3 million distinct IPs. The study found significant cross-swarm participation, with many peers participating in multiple swarms. The geolocational distribution of peers was also analyzed, with Australia, Europe, and North America playing a crucial role in influencing swarm size. The results of this study can aid in network usage prediction, bandwidth provisioning, and future network design, making it a valuable contribution to the field of network analysis and optimization.
Keywords: BitTorrent; cross-swarm peer participation; geolocational distribution; network analysis; bandwidth provisioning; network design; peer-to-peer file sharing; P2P networks
Abstract
Peer-to-Peer (P2P) file-sharing is becoming increasingly popular in recent years. In 2012, it was reported that P2P traffic consumed over 5,374 petabytes per month, which accounted for approximately 20.5% of consumer internet traffic. TV is the popular content type on The Pirate Bay (the world’s largest BitTorrent indexing website). In this paper, an analysis of the swarms of the most popular pirated TV shows is conducted. The purpose of this data gathering exercise is to enumerate the peer distribution at different geolocational levels, to measure the temporal trend of the swarm and to discover the amount of cross-swarm peer participation. Snapshots containing peer related information involved in the unauthorised distribution of this content were collected at a high frequency resulting in a more accurate landscape of the total involvement. The volume of data collected throughout the monitoring of the network exceeded 2 terabytes. The presented analysis and the results presented can aid in network usage prediction, bandwidth provisioning and future network design.
BibTeX
@inproceedings{scanlon2014analysis,
title={An analysis of BitTorrent cross-swarm peer participation and geolocational distribution},
author={Scanlon, Mark and Shen, Huijie},
booktitle={23rd International Conference on Computer Communication and Networks (ICCCN 2014)},
pages="1-6",
year=2014,
month=09,
organization={IEEE},
abstract="Peer-to-Peer (P2P) file-sharing is becoming increasingly popular in recent years. In 2012, it was reported that P2P traffic consumed over 5,374 petabytes per month, which accounted for approximately 20.5% of consumer internet traffic. TV is the popular content type on The Pirate Bay (the world’s largest BitTorrent indexing website). In this paper, an analysis of the swarms of the most popular pirated TV shows is conducted. The purpose of this data gathering exercise is to enumerate the peer distribution at different geolocational levels, to measure the temporal trend of the swarm and to discover the amount of cross-swarm peer participation. Snapshots containing peer related information involved in the unauthorised distribution of this content were collected at a high frequency resulting in a more accurate landscape of the total involvement. The volume of data collected throughout the monitoring of the network exceeded 2 terabytes. The presented analysis and the results presented can aid in network usage prediction, bandwidth provisioning and future network design.",
doi={10.1109/ICCCN.2014.6911846},
}