UK ISPs Resistant to Monitoring Users

ethericalzen writes "An article from BBC News online states that ISPs in the UK are resistant to the government's desires for monitoring their users' data. The government seeks to have ISPs turn off the access of users who are 'persistent pirates'. The ISPs are citing technical and legal reasons for why they do not wish to do this. Legals reasons include surveillance laws which prohibit ISPs from monitoring a user's data unless compelled by a warrant. Technical reasons include an inability to accurately identify copyrighted material that is legally being transferred over p2p clients, and copyrighted material that is being transferred illegally over p2p clients."

Comments

Didn't use the magic word!

[Chris Burke]

Legals reasons include surveillance laws which prohibit ISPs from monitoring a user's data unless compelled by a warrant.

Silly UK government! The secret password to get around the law isn't "piracy", it's "TERROR"!

I welcome this

[Malevolent Tester]

I'm glad the UK government is cracking down on file sharing. In particular, I'd like them to crack down on their own habit of sharing my personal information with every single bloated, inefficient, fuckwitted, semi-competent IT services provider who made a sufficient donation to the Labour party at the last election (Crapita, this means you)

Re:Yeah, right

[91degrees]

A final twenty million kicking each other to death for fake Burberry baseball caps.

They're actually kicking each other so that they can film it and put it on youTube. So they're all watching each other as well.

'persistent pirates' == everyone...

[Doug52392]

Those idiot lobbyists and corporate CEOs think ANYONE that uses BitTorrent is a pirate. But are they? Let's see, here's what I (at least attempt to because I have Comcast) use BitTorrent for:

1. Downloading large Linux install DVD images
2. Download legal, open-source programs
3. Download legally free files

The problem with this is that I bet NO ONE will actually sit there and read all the traffic logs. A computer will just flag customers who even so much as transfer a packet through a BitTorrent port as a 'persistent pirate' and cancel their service.

A computer can only say YES this person is using BitTorrent or NO he's not. The computer CAN NOT find out exactly what someone is downloading, and weather it's legal.

So if the UK wants to fall behind everyone in the Internet age and cancel EVERYONE out of the Internet, not much we can do but hope it doesn't happen.

Re:'persistent pirates' == everyone...

[bob.appleyard]

Deep packet sniffing can be employed, but that would be terribly costly. Considering the amount of traffic that a typical UK ISP would be dealing with, you're already talking about some massively parallel computer just to handle that. To then go and do some pattern matching stuff on every packet to see if it's a case of copyright infringement would not only be hell to actually code, it would take a ridiculous amount of computing power. What's worse, this could be defeated by encryption.

Some kind of shallow packet sniff could be done to check for hashes of copyrighted material, and ban people referring to that hash. That requires a database of infringing hashes, and new materials could be appearing all the time. It would also require knowledge of how every P2P protocol communicated these hashes. There is always the chance of mistaking some unrelated innocent communication with an illicit one, that will increase with the size of the copyright database and the number of protocols checked. And this can be defeated with encryption too.

There is another way of approaching this problem. Take a protocol, say, BitTorrent. You could use your database of copyrighted works, and check out trackers of those works for whether UK IP addresses are connected. You have a problem choosing which trackers to check, as you usually only get a name and a description. This one can't be beaten by encryption, but some elaborate series of proxies could elude it.

One of the common weaknesses with all of these approaches, beyond what has already been mentioned, is that the infringing works must first be identified. This could be done by downloading all the files and then manually checking them or by using some kind of fingerprinting technique. This is going to be either expensive in terms of equipment (big server farms for fingerprint analysis) or in terms of people checking the files manually.

The problem here isn't really technical, it's economic. Although most of the costs of blocking customers will have to be done by the ISPs, it will probably be up to the media companies to pay for identifying (say) trackers/hashes for copyrighted works. As this brief analysis seems to suggest that most of the cost is in the identification stage by a long margin, there would have to be a significant payout to justify this investment. I'm not convinced that it's there.

The other aspect of this is that the actual costs of P2P file sharing can only be guessed. Without at least some kind of checking described above it will remain that way. You don't know how many people are sharing this stuff, and you don't know how many of them would have bought it if they couldn't share the stuff. These are unknown risks, so deciding what kind of effort should be expended in avoiding them is a very very difficult task. It's hard to tell whether you've spent more money preventing the risk than the risk would have actually cost. Inevitably, they will have to choose whether to step off the ledge and actually do something on a big scale, or give up altogether. It seems to me that option with least risk business-wise would be to give up trying to stop it and use other means to reduce the probability of it occurring such as lowering prices or offering a better service than pirates.

But in practice, they're all for it

[mrsmiggs]

Bargain basement ISP Tiscali have already operated a similar scheme in cahoots with the BPI, but it's all fallen apart because Tiscali want the BPI to pay for the privilege of sending warnings and chucking people off. The most intriguing part is that the BPI are doing the investigation and instead of monitoring the packets of each connection they are monitoring the known torrents and connections to those torrents, which is clearly a far more practical idea than monitoring all packets. The Register have the full story