A new paper from the University of Washington department of Computer Science and Engineering, has investigated a problem with current DMCA notices and the methods used in dealing with them. It puts further pressure on anti-p2p groups like BayTSP to validate their claims.
Called “Challenges and Directions for Monitoring P2P File Sharing Networks ,or, Why My Printer Received a DMCA Takedown Notice”, the paper is a look into methods used to collect IP addresses for the sending of DMCA notices, and focuses on two main methods called ‘direct’ and ‘indirect’.
The paper shows that the tracking methods used by anti-piracy enforcement companies are not watertight, to say the least. Indeed, last year we already reported that it is possible to trap people into being reported to the MPAA or RIAA, by simply letting them click on the announce url of a BitTorrent tracker. The research from the University of Washington confirm these vulnerabilities, as they managed to receive hundreds of infringement notices addressed to a networked printer.
In August of 2007, data was being collected for a study of BitTorrent activity. During the experiment, the research team received over 200 DMCA complaints, despite never having violated any copyright. This alone should, and does, strike at the basic credibility of the organizations issuing such notices, and will be a big question if the content industries attempt to push for another ’3 strikes’ approach.
Then, in May 2008, the team decided to repeat the tests, first to see if things had changed, and to discover if they could later implicate other IP’s, ‘spoofing’ their presence. This time, there were almost 40% more DMCA notices, despite still not actually infringing copyright as claimed in the notices.
However, not only did the number of notices rise between the two monitored periods, the actual number of swarms monitored by the research team decreased. In August 2007 they had one notice for every 270 swarms. In April, that was up to less than one per hundred. However, without better knowledge of the torrents they were on each time, it’s hard to say if the difference was down to the choice of targets, or if anti-piracy efforts had been stepped up. For example, it could be that the increase was down to picking ‘hotter’ torrents, rather than an increase in enforcement.
Michael Piatek, one of the researchers involved in the project told TorrentFreak that they have contacted BayTSP and other enforcement agencies about the vulnerabilities in their tracking methods, but nothing has changed so far. A spokesman of BayTSP said he could not give a comment yet, but he assured us that their technical team will look into the research.
In addition, the researchers also checked to see if they could detect anti-p2p efforts, such as loggers. Whilst the paper gives one potential way to identify things, and later a method for using that to automatically create blocklists, the end result is that even with these major assumptions in favor of the blocklists (by eliminating residential IP addresses from consideration, for instance) at best, out of the 17 suspicious IPs found, only 10 were in such blocklists, 8 of them at colo facilities, tagged as ‘Mediasentry ‘or ‘Mediadefender’ (note that MediaDefender does not do enforcement)
In summary, the paper says that “potentially any Internet user is at risk for receiving DMCA takedown notices today. Whether a false positive sent to a user that has never even used BitTorrent or a truly infringing user that relies on incomplete IP blacklists, there is currently no way for anyone to wholly avoid the risk of complaints.” More dishearteningly for groups like Bluetack, however, is that it’s yet another kick at their claims of protection.
The last paragraph sums things up better than I can, however: “We have further demonstrated that IP blacklists, a standard method for avoiding systematic monitoring, are wholly ineffective given current identification techniques and provide only limited coverage of likely monitoring agents.”
We will give a more detailed breakdown of the study and its implications, soon.