[routing-wg]RIPE 52 routing-wg minutes
Joao Damas Joao_Damas at isc.org
Mon Sep 18 11:36:44 CEST 2006
With thanks to Arife Vural for taking the minutes, to my co-chair Rob Evans for making sure these got out there and to all who participated. See you soon at RIPE 53 or on the mailing lists. Joao Damas Routing-WG, Istanbul 2006 Chair: Joao Damas Date: 27 April, 2006 Afternoon Session - Previous minutes were approved. - Joao asked if there were any objections to Rob Evans being a co-chair of the WG. No objections were raised. - Philip Smith gave an update about Route Flap Damping obsoletion. http://www.ripe.net/ripe/meetings/ripe-52/presentations/ripe52- routing-deprecating-ripe-229.pdf The document, co-authored with Christian Panigl has been sent to the mailing list, but no comments have been received yet. Randy Bush: I read the document, liked it. I proposed vendors to change the implementation based on Geoff's presentation earlier this week. [http://www.ripe.net/ripe/meetings/ripe-52/presentations/ripe52- plenary-bgp-review.pdf] Gert Doering: I agree that the current way route-flap dampening is done is harmful. We should start work on a new recommendation how to get these extreme prefix noise dampened, but not 'normal' activity Joao: WG already has consensus, only thing is left whether the text reflects what we agreed. If so, let finish it. - Philip Smith talked about a Route Aggregation BCP He will have something to present next RIPE meeting about the recommendations. - Andrei Robachevsky gave an update about Authorization of "route:" objects in the DB. It's an update about the discussion that was in DB-WG last month. http://www.ripe.net/ripe/meetings/ripe-52/presentations/ripe52- routing-route-creation.pdf If the route-object holder change the provider, who will do the DB update? Ruediger: It's defined in RFC, and whois reflects that. Joao: We can change RFC and whois if it creates trouble. Ruediger Volk: Changing the procedure, I have not done the analysis of the consequences. Andrei Robachevsky: Need tweaking? Do we have to change paradigm? There is a work around, do best practice. Joao: Does anybody have any comments? I will summarize it, and to see the result. - David Ward gave a presentation about service peering and BGP for Inter-domain QoS Routing. [Slides weren't working during meeting, they are available at: http://www.ripe.net/ripe/meetings/ripe-52/presentations/ripe52- routing-bgp-qos-sicily.pdf] The proposal adds information on paths with varying QoS to inter-domain routing by adding it to BGP. Ruediger Volk: I'm not quite sure about the clarifications. This is for QoS for VPNS, does this apply IPv4? David Ward: Yes, it is. It is not tied to VPNs. Ruediger Volk: There is a scaling issue for BGP, did you take this into consideration while advertising same prefixes multiple times. David Ward: This is to be used between service sharing ISPs and not to be propagated globally through the Internet. Joao: Those messages are protocol specific? David Ward: Yes. - Ruediger presented some suggestions on how to proceed with Routing Security, IRR based authorization, Black-holing. Joao: Do you have a proposal on how to change the behaviour? Andrei: These data is available is the form of stats file. It does not tell the who is doing the authorization. Joao: Let's see what RIPE NCC can do about it. - Charter discussion: This was deferred due to lack of time at this meeting. - AOB There was no other business.