IETF-70 Routing Area Open Meeting ================================= http://rtg.ietf.org ADs: Ross Callon, Dave Ward Scribe: Deborah Brungard, Dimitri Papadimitriou Administriva and Area Status ---------------------------- Ross opened and reviewed the agenda. Routing Area Working Group Reports ---------------------------------- BFD (Dave W.): did not meet. Unfortunately drafts have expired (the author has not been able to update). Expect he will be able to now update and it will soon be done. Had discussion on P2MP but will hold off a bit for now before starting (need to resolve some remaining technical issues). CCAMP (Adrian): Meeting 2x. First session had a big bash on TE. I (Adrian) got fairly beaten up. We identified some areas as really application dependent. Starting discussion on Ethernet drafts (by Design Team). Second session we will be doing WDM, switching and constraints. FORCES (Dave W): Expect some new drafts will be moving forward. Need expert review on documents. IDR (Dave W): One presentation, Lou. Not going forward. ISIS (Dave W): Several new drafts. Tony Li will be completed soon. L1VPN (Dave W): Did not meet. In ccamp, a related draft on BGP discovery of (L1VPN) end points was discussed and decided may be more appropriate for work in softwires. MANET (Joe): Several items will be continuing. MPLS (Loa): Cleaning up backlog. Starting several new documents. Security framework draft on GMPLS and MPLS most important. If people interested from the routing area and have time to review, welcome that. Multicast will be starting on Monday a Last Call. OSPF (Ray): Several drafts in progress. Multiple instances, multi-segment PW. OSPF lite felt should be a BCP and will progress in that way. PCE (Adrian): Doing well. PCE protocol at v9. Quite stable. 9 implementations, quite a few inteop. Starting to look at widing the applications, P2MP, TE. PIM (Mike McBride): couple of documents updating. RPSEC (Dave W): Not meeting this time. Ready to last call BGP Security Requirements. RTGWG (John Scudder): Not meeting. Their planned presentation will be given here. GTSM has been published as a proposed standard. Framework is ready for last call. SIDR (Sandy): Met. Couple of updates on-going on drafts. Two new issues on composite structures. Huston proposal is not going to be progressed. VRRP (Dave W): did not meet. Dave: FYI, BOF tomorrow on routing over lossy networks for anyone interested (overlaps with ISIS). Quick announcement: SAAG (Ross) ------------------------------- SAAG will be preparing for a BOF in Philadelphia on key management for TCP authentication and other router integrity protection exchanges. Folks interested in Routing may want to attend tomorrow. A thought on Routing, Addressing, Scaling (Ross) ------------------------------------------------ Issues: 1. Routing table size (router FIBs and convergence) 2. Availability of IPv4 address allocations Current work: 1. Radir statement (draft-narten-radir-problem-statement-01.txt) 2. LISP and related work (RRG) 3. BGP enhancements (IDR, et al) 4. Larger FIBs and software updates (vendors) 5. IPv6 (some deployment) Question to community: 1. Is other work needed? Eg1: Is anything needed to make IPv6 deployment easier Eg2: Should there be a way to push back on announcement of excessive prefixes? What does excessive mean? How to Push?? Please feel free to bring forward ideas. Discussion: Dave M: wanted to say its more economical issues than technical. And asking, is IETF where the work should be done? Wondering what you (ADs) think? Dave W: if there are things which we need for interop, then we can try as ADs to make it happen. What Ross is saying, IF there is more work, we can make it happen. T-MPLS Update (Stewart) --------------------- Stewart presented slides. A proposed joint working team of ITU and IETF will be established to progress work. Planning to liaison to SG13 to recommend not approving two OAM Recommendations (G8113 and G8114) as concern on use of Label 14 and inconsistency with RFC3429. Discussion: No Questions. BMWG and routing protocol dependencies (Al Morton) -------------------------------------------------- Al presented slides overviewing the BMWG. Dave: how we got started on this was related to convergence. We had several drafts with discussion at IESG. Now going to recommend that when the BMWG does a draft on one particular protocol to have a tighter relationship with the working group doing the protocol. Please get more involved, it is critical work as its what our protocols get measured against. So asking Al to get in touch with the wg responsible to let them know when work is being done. Performance Simulations for IPFRR (Mike) --------------------------------------- Mike presented slides. Stewart another thing of interest to analyze is the FIB size change. Mike: depends if want full coverage. Stewart: depends if just looking at the link or LSAs. So be interested to know on the expansion on the FIB size. Mike: we have looked at FIB size. Not sure if looked at that though. Ross: what was the network topology used for the simulation? Mike: its an experimental network (GEANT, an EU research network) Ross: do simulations also work for a larger network? Mike: should;-) Ruediger: another aspect is the metrics, are the metrics used what actually used in a network? Mike: not sure if actually, would like to use more realistic, but not necessarily available. Dave W: is your modeling program avail as free software for others to try with their LFAs? Mike: no its a hack;-) Stewart: its common in academic networks to use hop count, which is different from real networks. Should try. Though this is not what you have done. Alia: there are some research networks willing to share, I will see what I can share with you. Routing Change Process ----------------------- Dave W. I have been approached by several chairs concerned that there are a number of Modifications being proposed by other standards groups and within the IETF itself. Its not going well, as its sort of an adhoc process for how we do it. This has been an ongoing process for several years. Often no review until last call is underway. Stewart: agree, how do you want to do this? Dave W: MPLS/GMPLS has a change control process, which we may want to consider to use. Its not exactly applicable, as we want to solve that internally groups work close together. MPLS process requires a full fledged requirements document. Here, say that its up to the chairs if a requirements document is needed or if the requirements can be within another document. So what we will say that any request to change a routing protocol, goes to the routing working group responsible. And be looking at having a formalized change process document. Do people feel this is needed and should it be formalized? Stewart: yes, yes John Scudder: yes, yes Acee: yes, yes, do not want to double the number of documents, maybe can have a section in that groups document. Dave W.: say its for the chairs to decide how they want to do. George: as someone usually opposed to process, on this, in favor. Kireeti: whats to prevent the SDO that comes to us, writing requirements, and us saying no and they do anyway. Dave: reference the IANA process. Kireeti: I wrote that for OSPF, you will need to document for other protocols. Ross: Its a bit different if another standards body does it vs. internal in IETF. Kireeti: I agree. Ross: we need to say if something comes up that we feel is dangerous to our protocols, we need to try and work with them to fix the problem. And if can not be fixed, we need to speak up in the public and say it. Dave W: best we can do is set up an allocation of code points. Alex: I see a need for clarification of inter-SDOs and also clarifying how when one area within IETF proposes and how to review, not sure need more red tape, do not want that. I do not want that we need requirements documents and framework for everything we want to do. Its too much red tape. Dave W.: its for the WG chairs to decide what they need to understand. I also do not want more red tape. George: When a WG chair puts a document to IESG, probably should have a section on what codepoints are allocated in that document, and then ask that WG chair if they agree. Kireeti: how well is the early allocation and experimental allocation doing. (Several said using) So if other SDOs ask for them, do we know how to respond to them? Dave W: I do not know, we should. Stewart: I think its the WG chair that wants the allocation, and not necessarily the WG that owns the protocol, may want to look into. Dave W: so no one opposed to us doing a draft to formalize, we will do for the next meeting. Open Mike --------- No discussion.