The BitTorrent Ecosystem includes millions of BitTorrent peers, hundreds of active trackers, and dozens of independently-operated torrent discovery sites. The Ecosystem is further fueled with distributed trackers using Distributed Hash Table (DHT) and Peer Exchange (PEX) functionality. The Ecosystem also includes ?interdiction companies," which attempt to curtail the distribution of targeted content.
Despite its importance, both in terms of its footprint in the Internet and the influence it has on emerging P2P applications, the BitTorrent Ecosystem is only partially understood today. Many communities (including P2P researchers and developers, ISP researchers and engineers, copyright holders, and law enforcement agencies) would like to have a comprehensive and in-depth understanding of the BitTorrent Ecosystem, as well as tools for mapping the Ecosystem in the future.
In this context, the PI and his graduate students are exploring two inter-related research directions. First, they are developing tools and methodologies for comprehensive exploration and mapping of the entire BitTorrent Ecosystem. Second, they are examining of how the Ecosystem can be attacked and defended.
The expected results for the mapping research include: new public-domain tools and methodologies for mapping and analyzing the Ecosystem; a comprehensive mapping data set, which will be more than an order of magnitude larger than any existing data set and will essentially cover all trackers and peers in the public Ecosystem; novel estimation methodologies based on importance sampling, incorporating measurement samples from both centralized and distributed trackers.
The expected results for the BitTorrent attack/defense research include: measurement and evaluation methodologies of ongoing attacks from ?interdiction companies"; an in-depth study of the seed attack, whereby the attackers attempt to prevent the initial seed from distributing the file into the Ecosystem; machine learning algorithms for defending against the pollution attack in BitTorrent; and tractable deterministic and stochastic models for the dynamics of BitTorrent attacks, providing critical insight into BitTorrent vulnerabilities.
