P2P Throttling To Get Worse ?

[GhostShip]

Many of you have heard of the next company as a provider of equipment that slows down p2p network traffic, it seems a few tweaks have been added.

http://www.tmcnet.com/usubmit/-ellacoya-networks-bolsters-deep-packet-inspection-capabilities-/2006/10/11/1973373.htm

Ellacoya Networks, Inc., a leading provider of carrier-grade Deep Packet Inspection (DPI) solutions for broadband IP networks, today announced it has enhanced the DPI capabilities of the Ellacoya IP Service Control System with its new 6.4 Software Release. The new release adds application detection and classification functionality for encrypted applications, including bandwidth intensive encrypted peer-to-peer (P2P) such as BitTorrent and Skype (primarily in North America/Europe) and Winny2P and Share (primarily in the AsiaPac markets).

The networks mentioned above have so far enjoyed being able to bypass ellecoya traffic limiting/shaping by being encrypted. I'm sure a few adjustments to the protocol will turn the balance back in favour of giving the consumer what they have paid for.

[Neutron]

They should use some standard encryption like SSL, so the lame ISPs who charge you and don't provide the service can't distinguish p2p traffic from other traffic.
Just because the protocol is encrypted it doesn't mean they can't see what it is. They will know it is a bittorrent handshake even though they can't see the encrypted data that is being sent. SSL is something used by many protocols, like HTTPS. They have no way of knowing what traffic is passing through the connection. But even then, bittorrent isn't only used for "illegal p2p" as the anti-p2p organisations would define it, it actually has legitimate purposes. Lots of companies release new large files through bittorrent to save bandwidth. Blizzard use bittorrent to distribute the new patches for world of warcraft as they are released, because that is the best way to distribute it to millions of players at the same time. I am sure more companies will use bittorrent or similar protocols in the future to distribute their large files. ISPs have no right to throttle or block their customers' bandwidth when that's what the customers pay them for. throttling bittorrent traffic is just as bad as throttling HTTP traffic.

[7]

I don't think so many ISPs are too concerned about what users are doing to consume so much bandwidth (exchanging copywritten information for instance), rather consumption overall is the issue as many (if not all?) service providers price their products as low as possible in order to compete against other providers based on knowledge that most subscribers won't actually utilise their connections to the fullest.  As bandwidth usage patterns increase ISPs are forced to either implement restrictions, increase subscription price, both, or face potentially losing business to competitors who do chose to restrict the heavier users to allow them to continue to charge less, attract more subscribers, and so on...

I don't think use of 'SSL' would change much overall as it doesn't take into account the large number of unique connections peer-to-peer clients tend to establish when operating which can make it quite obvious a subscriber is more-than-likely running a peer-to-peer service and not some simple SSL-like link out to another machine or two.  Encrypted BitTorrent could be relatively easy to identify for at least these reasons (not even taking into account typical ports seen used in communications (common BitTorrent client ports) and associated unencrypted services like DHT and such).

I think it could prove difficult to circumvent ISP control in the long run given the unique position they're in to be able to govern the system.  Personally I'd like to see some sort of open wireless mesh network arise (or something?) as a sub-network of sorts (to work outside the bounds of the typical Internet) to bring real control into the hands of the end-users and not money interests (which have a tendency to fuck everyone/everything over for profit/survival).

Just another battle to add to the mix of conflicting interests in this world I guess.

[KM]

Most people would assume that it would benefit larger ISPs more to "Traffic Shape" than smaller ones (as if it cuts down say 30% of traffic - that is more money to a larger ISP) however in reality small ones have smaller profits so are more likely to need every penny they can, but also most ISPs the majority of their traffic is "free" (as in they only pay for the physical equipment) whereas smaller ISPs typically have to use transit providers for most of their traffic

My ISP for example (Telewest) i have found that around 60-80% of typical p2p destinations (ie. normal home connection ISPs in europe and the US) it is reached directly using peering (free bandwidth), whereas most datacenters it goes over transit providers (which costs money) - therefore p2p is almost free to them, certainly better than downloading things from central servers anyway

A small ISP however will probably have all of their traffic going over transit connections (they just get say 2 connections from transit providers for redundancy), meaning all of that bandwidth costs money - at best they might have a connection to the London Internet Exchange, meaning they might get free bandwidth to other ISPs there, which might give them free traffic to most UK based destinations and a few european ones, but they will be paying for a much larger portion of their bandwidth usage

Of course most ADSL providers (Certainly in the UK, and also in many other countries) have to pay the phone company large amounts based on bandwidth usage, for those providers it is often the case that the single largest cost they have is the cost the phone company charge them to get the bandwidth from the users home to the ISPs network. In those cases if they can reduce the bandwidth usage of the users connection it saves them a lot of money, which of course makes the cost of a small box for messing up p2p traffic seem much more attractive.

Of course we won't tell them that the expensive equipment they all use for p2p shaping isn't actually needed and an old computer they were going to throw away running linux would be more than able to do the same job with hardly any cost (although some of the larger ISPs might need a couple of medium spec computers instead)

[7]

Of course most ADSL providers (Certainly in the UK, and also in many other countries) have to pay the phone company large amounts based on bandwidth usage

I don't believe it's so much based on charging for bandwidth consumed these days as, I believe, most U.K. non-LLU DSL ISPs (whew) have moved over to B.T.'s Capacity-Based Charging (CBC) method for their 'Central' pipes now which charges a fixed amount for the link irrespective of bandwidth usage, as opposed to the older Usage-Based Charging (UBC) method that used to be more common.  So now the incentive for ISPs to reduce bandwidth consumption is being based more around the idea that the less bandwidth the average subscriber consumes, the more people they can support on a single B.T. 'Central' pipe.  The CBC pricing method pushes ISPs to utilise the Centrals fully by packing as many subscibers onto the link as possible because a chock-full Central still costs the same amount as an empty one.  Problem though is as subscriber bandwidth usage increases with the arrival of more-and-more bandwidth-heavy content (video seems to be taking off in a big way these days) ISPs are starting to run into some serious problems managing bandwidth, particularly during peak-time hours, across these lines.

Nice description of B.T.'s network here for anyone interested.

[GhostShip]

That link seems not healthy 7 

I took a peek around and this one looked good with link to a similar named item in the first few lines.

http://kitz.co.uk/adsl/BTCentrals.htm

http://www.kitz.co.uk/adsl/equip2.htm

Cheers mate 

[7]

Sorry about that I had a quotation mark in there. Fixed.

It's a good site, I learnt a lot from there (and the bods at ADSLGuide.org.uk).

Regards.