Scraping BitTorrent Trackers

State: completed by Gian Marco Jutz

For evaluation of BitTorrent related projects real life data is of great value. Data about the number (and maybe addresses) of peers that participated in a certain torrent at a certain time would be an excellent basis for evaluation and validation. BitTorrent trackers offer an interface to acquire this information, called scrape. This thesis will focus on development of a tool to collect data from a tracker in specified time intervals (e.g. 15 minutes). Furthermore, the tool will be deployed distributedly to avoid IP-Bans and collect real life data about torrents and swarm size.

10% Design, 60%-80% Implementation & Deployment, 20%-40% Documentation
Java programming, Linux, Database skills (P2P knowledge of advantage)

Supervisors: Dr. Thomas Bocek, Andri Lareida

back to the main page