On Thu, 31 Oct 2002, Henk Uijterwaal (RIPE-NCC) wrote:
Folks,
I've put the latest version of the Active Measurements BCP Internet Draft, that I mentioned during yesterday's Measurements panel, online at:
http://www.ripe.net/home/henk/draft-ietf-ippm-owmetric-as-01.txt
I think this draft makes considerable sense and addresses some of the measurements which might be used to characterize performance (as the title of the WG says) which is good. I'm also interested in provoking dicussion about the active measurement requirements for the other purposes which were touched on by the panel. Primarily, collecting traffic information for purposes of deriving traffic matrices, capacity projection, traffic engineering and attack profiling and detection. To put a fine point on it as I tried to on the panel: Can we come up with some best current practices for addressing these needs and deliver them as a clear set of requirements, both high- and low-level to equipment vendors? In addition to the very specific stuff of how to sample traffic or how not to, I'm thinking a framework document, applicability statement, or the like which lays out the big picture of how the protocols are envisioned to be used; what real-world problem(s) they solve. I chatted with Jen Rexford briefly afterward and she said that the PSAMP WG has a draft framework doc which appears to be this: http://ietf.org/internet-drafts/draft-ietf-psamp-framework-00.txt The IPPM has a few including some which have become RFCs. What do people think about how well they meet the needs they have at hand? Would some even higher-level description of an activity such as capacity planning be useful where we could then crystalize some tools we need to get that job done? What about laying out exactly how we think they need to be used so that if some part of the requirement changes down the road, the relevant adaptations can be made? Talk amongst yourselves. Tony