Quote:
Originally Posted by Paul Delaney
Can somebody please confirm that the following is correct (want to try a different tack with my MP)
|
DPI can already do all that you do not want it to do. There are 2 parts to traffic: the header and the visible content.
For traffic shaping, only the header data is read. This says what type of data it is. Video and music have different header data from standard text on a web page so it is relatively easy to give each a different contention ratio. That is a valid business purpose to improve the service.
If you don't like your ISP limiting you because they have oversold the capacity of the pipes they use, move to an ISP who has invested in their own pipes and invests again before traffic congestion begins to effect performance experienced by their customers (costs more than £10 a month).
The content data is also available to the DPI system. Processing that too costs a lot of money and CPU power, so there is no financial reason to include that additional data for performance reasons.
However, there is nothing stopping anyone using DPI to copy everything. It 'sits on' the data stream and can view whatever it is programmed to view. If a court order walked in the door, I suspect that any UK ISP already installs all the equipment necessary to stream off 100% of the traffic data for a specific data stream. It just needs the court order to do so.
What Phorm is doing is taking that 100% data stream and running its own software on it. When it arrives at the Phorm supplied software, all the PI data is part of the stream. The IP address (that is usually classed as PI) is a problem for the Phorm system as the ISPs give out random IP addresses each time a person connects their router to the internet, so using IP address as idenifier would not be feasible for Phorm. This is why they give everyone a 'random' UID so that their surfing round the internet can be followed. This is much more personal than an IP address and each browser used on the computer will be given a unique UID. If a building shares an IP address - even it the IP address is static, there could be hundreds of people using the address. By identifying the actual browser being used, there is more chance that the UID identifies the same person for each visit to the internet.
The way the system works, the UID, once set, is available in every data stream sent from the user's browser so the system can immediately add the current data collected to that already stored by the system.
I know that BT/Phorm say that some data like https and blacklisted sites is excluded from being profiled. They do not say that it is not first intercepted so that it can be identified as data that should not be profiled. Phorm do say that they profile URLs and nowhere do they say that they differentiate between http and https URL content - only the content on the https web page is excluded. For this reason, I used 100% in the above text.
---------- Post added at 02:00 ---------- Previous post was at 01:50 ----------
Legally, a minor may not undertake a contract. In the advertising world, teen is about 9+ (you only have to look at who reads 'teen' magazines).