100 Publishers Create Most of BitTorrent Content

Millions of people use BitTorrent, the peer-to-peer file sharing network, to download content, but only a handful of its users are responsible for sharing most of the content, new study finds.

Peer-to-peer file sharing is a name for networks that allow users to share files , such as computer programs, multi-media (audio, video), documents, or electronic books with each other directly, without the involvement of a central server. Users can share, or “seed,” files on their computers for others to download.

A team of researchers from Spain, Germany and the United States studied 55,000 streams of content shared by more than 35 million individual computers. They found that sharing on top torrent sites is far more concentrated than was previously thought, with just a few publishers working to provide the bulk of both legal and illegal content.

In fact, the researchers found that just 100 publishers were responsible for 67 percent of all the published content and 75 percent of all downloaded content. In addition, 30 percent of all that content was offered by “fake publishers” -- media companies looking to poison the file- sharing hubs, or malware authors pushing malicious downloads.

These malicious and fake media files – which are intended to frustrate piracy or to spread malware – make up around a quarter of all download activity, which translates into millions of downloaded files, the researchers found.

Of the slice of download activity from legitimate publishers, the researchers identified two main sources: for-profit publishers who are solely interested in driving traffic to their websites and so-called "altruistic" top publishers, who share the content they have and consume content for their own personal use.

Sharing linked for-profit publishers accounted for roughly 30 percent of published content and 40 percent of downloading activity done through BitTorrent. On the other hand, the altruistic top publishers were responsible for around 33 percent of published content and 25 percent of downloads.

The researchers had a relatively easy time pinpointing fake and for-profit top publishers because they tended to publish large amounts of content and used multiple account names that were linked to the same IP address, while rarely downloading content themselves.  On the opposite end, altruistic top publishers offered a small number of files, while downloading and seeding content in equal measure.

The research was published at the ACM CoNEXT conference this past December and is available for download here.

• Seven Modern-Day Technologies Sparking Controversy
• 10 Things You Must Know About Malware Infections
• File Sharing Software Reveal Users' Private Information