On 29 apr 2014, at 16:26, Gordon Lennox <gordon.lennox.13@gmail.com> wrote:
I know it is late but BEREC has just had a public consultation on:
"Monitoring quality of services in the context of network neutrality"
based on a draft report, a draft which I think is still worth reading.
See:
http://berec.europa.eu/eng/news_consultations/ongoing_public_consultations/2...
I have not had energy to engage...because too many people that have interest in these kind of things and write documents do not know enough how Internet works, how measurement software works today, what is wrong with todays software and... Anyway, I have now looked at this....that is a lot of words...and it tried to dive into some of the issues with detecting how the network behaves. The problem is that they try once again to write text that everyone should understand, but because of that they loose the precision that is needed for a document like this. I do find some pieces that are interesting, for example this sentence:
When measuring IP layer metrics, the transport layer protocol (typically TCP or UDP) and application layer protocol (e.g. HTTP) are relevant for the measurement methodology.
That should have been 25 of the 50 pages just on that issue. But no, it was one sentence. Another key is this:
Furthermore, deciding the set of measurements to run is beyond the scope of LMAP, and has been left to the organisation which will manage the measurement system once deployed. In this way, the LMAP deliverables will develop an open and flexible architecture, with a likelihood of supporting the needs of the NRAs.
Which is correct, but what they miss is that measurement mechanisms is already a business for commercial companies (and of course individuals and others that are married to their ideas... ;-) ). So... I think the only solution is to: 1. Look at passive monitoring (that is indeed described in this document) 2. Continue to ensure the specification ISPs give matches what their customers do believe is what they get (and if their customers feel fooled, they do, and having ISPs following some ITU Y.something standard does not help) And this should have been the 2nd 25 pages of the report. That said, I think they correctly do reference various IETF working groups and explain why their work can not be used as of today, while that there might be some hope. I would not hold my breath though. Patrik