NNSquad - Network Neutrality Squad
[ NNSquad ] Re: Speculation, how AT&T can implement "copyright filtering" without wiretapping/dpi...
I think there are a few misunderstandings... On Jan 28, 2008 9:07 AM, Lauren Weinstein <lauren@vortex.com> wrote: > The only reason that "offending" torrents can be so easily identified > (especially in an automated manner) is that the operators of such > systems have not seen a need (up to now) to obfuscate this level of > information. And Eric wrote > 2) What happens when (inevitably, in response) technology comes into > common use that obfuscates the content? Again, based on past > experience, my guess would be that the MPAA et al will decide that > any content they suspect and can't examine/identify could be > objectionable and should be blocked. No. This is rather a fundamental limitation of open-world piracy and the content being distributed. Remember, the content being distributed that the ISPs care about is almost entirely video content and large pirated programs, which must render to the human eyes or computers, and it is "open world" piracy where it is easy for people to get to the content, which causes it to be a high volume issue. Thus all the schemes for "video DNA" and "audio DNA" tracking are robust in the arms race, at least with a bit of effort. If the transformations to avoid detection require inserting a huge amount of noise and disruption, the MPAA and ISPs wins, because that shuts down piracy just as sure as blocking the content. Just as you aren't going to read a trashy romance novel about black footed ferretsif its piped through a CAPTCHA program, you aren't going to want to watch "Untraceable" through the video equivalent. A similar arms race occurs in the software side, as portions of the code could be executed and compared for functionality. (This is harder, as it treads dangerously close to the halting problem, but if the pirates have to dynamically obsfucate every independent torrent of Microsoft Office, its going to screw up their day. Anyway, its not like Microsoft is not already effective at playing Whak-A-Mole using the BSA hammer) Accept it, identifying that a LARGE piece of content is a blatent copyright violation is a straightforward technical problem, and robust to the arms race with just a little effort. It can also be very robust to false positives. We aren't talking small content here, but 20+ minutes of video, or Gigabyte programs, in a single sequence. If anything, such a mechanism is far less prone to abuse than DMCA takedown letters. ACL-injection must be short lived (ACL capacity in routers is limited, so you only want to block the graph as it matters) and necessarily pairwise (to avoid having to do deep packet inspection or cause big false blocks), so it can be done very selectively. It also requires preestablished relationships (where much of the DMCA takedown problems come from "someone out in the blue") The biggest problem is actually not attacks on fair use, but screwups where an umbrella organization believes it has the right/duty to protect copyright on the file but the copyright owner doesn't (which happened to Cory Doctorow). For video, this seems a significantly rarer case, because of the barrier to entry and the probable commercial nature of any enforcement service. Likewise, as Richard just mentioned, accept that FINDING the participants is easy. This is "Open world, P2P piracy", which must, by definition, make it easy for people to find content and find the sources hosting the content. The alternatives for pirating data, "Closed world" piracy, solves 95% of the problem for the MPAA and ISPs anyway, as if people can't find it, they can't pirate it. As "closed world" groups open up, then it because reasonable for the MPAA to infiltrate the world and shut it down. The other alternative, "Open world, point-to-point" piracy is not effective for these large files. It works for music, but MP3 encoded music files are much smaller than video files. You have a very hard case for neutrality to make when the opposition is: a) It works with VERY high precision and VERY low side effects b) It does not affect non-infringing uses of BitTorrent. c) It cuts the Internet bandwidth used by a huge fraction. d) It does NOT involve wiretapping/deep packet inspection, or even require the ISP to make a determination at all, but simply allows copyright holders to enforce their rights. e) They lie and also say it can also stop kiddie porn. At the same time, wouldn't you much rather have a system of "Spider and ACL-block" over a DPI scheme?