Torrent Recommendation System Based on Data Gathered from the Mainline DHT

BitTorrent is the most widely used peer-to-peer file sharing protocol and attracts millions of users every day. It allows for easy distribution of all sorts of files including movies, TV shows, music, games, and other applications. In this thesis a closer look is taken at what files users are sharing. To acquire the necessary information a list of active torrents and in particular their infohashes is extracted from a popular torrent indexing website. In a second step all peers sharing a given torrent are found using look-ups in the Mainline DHT and aggregated in a database. After gathering enough data torrent recommendations are calculated using the technique of collaborative filtering. These are then presented to visitors on a website.