As part of the research into improving long term BitTorrent performance, the Tribler group is conducting research into swarm selection.
One aspect of this research is now converted into a uTorrent application.
This application allows users to define the number of torrents they want to seed and let our state of the art algorithms decide seeding which torrents improves the overall BitTorrent community the most.
As this is ongoing research, improvements in the algorithms should occur. Eventually improving Quality in whole the BitTorrent ecosystem, in which Quality could be described as the number of fake swarm present in the system.
For more information see SwarmCacheManager.
To provide a user with the best advice possible all health levels of a swarm are gathered by the app and used in our calculations.
Specifically we use the swarm age, #seeders and #leechers info to determine the health of a swarm over time.
This information is stored on our secured Linux servers at a non-profit organisation: Delft University of Technology.
Similarly to standard web servers, the logfiles store the IP addresses of computers that access it.
These IP addresses are modified such that we only store the first 3 octets (ie 188.8.131.52 will be stored as 123.123.123.xxx).
The server logs may be retained for the duration of the research project.
This trial will not record personal data, psychological information, zodiacal sign, nor other sensitive data of users.
No Person-identifiable data such as names, addresses or contact details are recorded.
Whenever a new advice is requested from the server.
The following data is stored in our web server logs: IP address (modified as described), nr of files wanting to seed.
For each torrent: infohash, trackers, #seeders, #leechers, time seeding, total downloaded, total uploaded, availability. At no point will we make this information public.
We reserve the right in the interest of scientific progress and
advancement of Open Source P2P software to make aggregated P2P
performance statistics and anonymised records available from these
trials to the public. We will never publish any personally
identifiable information in any form or function.