The increasing use of peer to peer software is creating more than just a piracy problem. The use of UDP by P2P can generate new network management and legal issues for ISPs.
In recent times, popular peer to peer software has started to use the UDP protocol as opposed to the more common TCP protocol.
What is UDP ?
UDP means ‘User Datagram Protocol’.
This protocol was designed to send short packets of data and is renowned for its speed. It is typically used in the operation of the DNS (domain name system). Unlike TCP protocol which has built in error checking and sends data in an order and without duplication, UDP does not have these characteristics.
Some Peer to Peer software (such as Emule) is now using UDP due to its speed. The use of UDP causes problems for ISPs because it does not support error checking. So requested packets of data may not actually be received by the computer requesting that data.
What is the problem for ISPs ?
Among other things the absence of error checking results in a situation where, for example, a customer using a 1.5mbps DSL service downloading data using UDP-enabled peer to peer software may request 6 gigabytes of data. The service cannot obviously download all of that data quickly due to the bottle neck at the DSL service.
Although the customer may not be able to pull down all of the data, the the data still responds to the request, so it passes across the ISP’s network boundary through a wholesale network (such as Telstra’s).
As the data has passed over the ISPs network, the ISP is charged for that whole of the 6 gigabytes of data, even though the customer may have only downloaded a fraction of that across the ‘last mile’ DSL tail. Most ISPs’ systems can capture this data for billing purposes, but many don’t have effective T&Cs in place that allow them to charge for this data.
A customer in a TIO complaint or legal action could argue that they never downloaded or used the 6 gigabytes of data, and they could prove this if they counted the data that entered their network.
What should ISPs do ?
ISPs should carefully review their terms and conditions to ensure that their terms unambiguously allow them to bill for data requested by the customer, that passes across their network and across the network boundary, but which may not necessarily reach the customer or their service.
It is equally important that ISPs notify their customers of the risks of peer to peer software and the potential for them to incur significant excess charges on their bill. Effective education and resources for customer use, could prove pivotal in defeating a TIO complaint.
Loading ...
Bah – no carrier or ISP I’ve ever worked with have ever defined data that hits the BPG vs. the data that hit the modem. Why because the lawyers who write up terms and conditions barely understand what an IP address is let alone the complexities of transmission and BGP.
As a result in the over 9 different ISP’s I’ve worked for the “billing and charging” sections of the agreements have always state “we’ll give you an invoice and you’ll pay it”. They have never gotten into the specifics and in all my years of doing usage disputes (from the average punter to very large wholesale customers) no one has been issued a wound up notice for not paying a usage bill on the basis outlined in your article. Essentially the issue has never been challenged.
UDP is a problem but ISP and carriers really don’t care since anything that hits your IP block, be it one of the usable IP’s or the broadcast, is simply billed to the customer, irrespective of whether the packet was even delivered to a operating service.
For example its quite common for customers who are getting layer 2 bridged DSL services to have their ISP assigned IP block ready to go before long before the DSL is provisioned.
I had one customer whose provisioning took 30 days after we’d setup the account and had received a bill for about 3Gb of usage. It turned out their IP’s had been hammered by 100Mb a day of UDP port scanning traffic caused by a particularly virulent period of worm activity on the net.
As I explained to the customer, a DSL connection is like being asked to build a road to your house. Sure your driveway might not be ready but we still had to have the road built and if tries to get to your house then that is still considered a chargeable event.
Everynow and then then, especially when there is a provisioning delay the usage might be credited (but we’re talking about a couple of bucks) but at the end of the day usage is not what you’ve requested or even what has been delivered to your connection. Its what has hit our BGP/core router.