IETF-66 Routing Area Open Meeting ================================= http://rtg.ietf.org Administriva and Area Status ---------------------------- Geoff Huston and Bruno Rijsman agreed to take the minutes of the routing area open meeting. Bill Fenner stepped down as RTGWG chair; John Scudder will take on the role of co-chair of this working group. There is the intent to revive the routing area directorate to perform timely reviews of routing area documents. It is anticipated that the directorate members will be a resource for WG chairs to undertake early review of routing drafts prior to the final steps of IESG submission. The AD review will also use the routing area directorate for comments as part of the AD review process. The ADs intend to streamline the document review process to reduce the queue size and resulting delay. Routing Area Working Group Reports ---------------------------------- BFD - Not meeting this IETF. Will last call the complete document set. Upcoming draft on BFD for multicast. The directorate should be involved in review at this stage CCAMP - New co-chair Deborah Brungard succeeds Kireeti Kompella. There is steady progress on drafts. Current focus is on IGPs, TE capabilities, automesh, ASON routing, and MIBs. The GELS BOF relating to GMPLS control of ethernet environments will be folded into CCAMP, and GMPLS control protocols will be used. In this proposed direction CCAMP will not be involved in the definition of the dataplane (this will be left to other standards bodies such as the IEEE), which is addressing one of the more contentious issues that surfaced in the GELS BOF. FORCES - WG chairs not present to report to the area WG. IDR - There is progress in terms of moving documents through the IESG. A number of RFCs have been published after IETF 65 and a number of additional documents are to be passed to the IESG in the coming weeks. There are various efforts to add things to BGP outside of the routing area. The softwire WG was called out as one particular example but there are others as well. In the particular example of the softwire WG it was agreed to closely coordinate the work with all relevant working groups (IDR, L3VPN, and L2VPN) including cross-posting of the documents on the relevant mailing lists. The meta-question is how to avoid duplication of effort, having multiple solutions to the same problem, and re- inventing the wheel when one WG is working on extension to protocols in another WG. Yakov suggested that this is a management issue. Some further discussion with the ADs was proposed. Other areas and other WGs do work on extensions to routing protocols, with the risk of inadequate management of the outcomes to ensure coherence and consistency of the extensions. This is asserted to be a matter for AD attention and work management. Some coordination effort across WGs should be undertaken as early as possible when work on routing protocols is taken up in other working groups ("early cross-area review" is the procedure being invoked here). ISIS - Some further work items (multi-topology ISIS and extending the LSP space) and rechartering are underway. Implementation report requirements was noted as an area for discussion. L1VPN - Progressing well. Framework draft close to completion and use specification underway. Some drafts affect BGP or OSPF; there has been some cross-WG and cross-area review. MANET - The 3 core documents have been updated. The WG authors are working on a common messaging format and the corresponding document has been updated. Work on a common neighborhood discovery work is underway. There are at present pro-active and reactive approaches and some effort to pull these together. The chairs of autoconf and MANET are working on a common architecture document and will publish this post-IETF. This document would benefit from early cross-area review. MPLS - A lengthy agenda at this IETF. The highlights are point-to-multipoint TE LPSs and a report of the meeting with ITU-T Q12-SG15 on GMPLS. OSPF - All the OSPF WG documents have been completed at this stage (MIB on v2 and security on v3). There is also the process of re-charting with MTR and OSPF-MANET are under consideration. Some active discussion on flooding optimization. Some OSPF WG review of the CCAMP documents was called for. A respin of v3 close to last call; the hope was expressed that v3 will have fewer respins than v2. PCE - 3 documents in the RFC Editor queue. The base protocol space is now stable, and a call for review was made. There have been some proposals for further work, and the chair would like to hold off on further spec requirements until there has been some experience with the base specification. There was consideration of an experimental track on manageability of the PCE specifications. Policy work is outstanding as is consideration of the complete requirement set. It was remarked that there might be an implementation of the base protocol. RPSEC - Finishing up with work on generic threats; is back in the RFC editor queue. OSPF vulnerabilities and BGP attack tree documents are being revised and appear to be close to completion. BGP security requirements is also considered to be ready for last call. Without further new items (e.g. additional routing protocols in need of a security analysis) the WG will have completed its current charter. RAWG - Not meeting this IETF. Due to some recent new proposals there is now an open question about loop free / microloop detection which needs to be resolved prior to last call. Advanced work can progress after the basic work has been finished. RTGWG - Not meeting this IETF. Alex Zinin reported that due to some recent new proposals there is now an open question about loop free / microloop detection which needs to be resolved prior to last call. Stewart Bryant said that there was an outstanding issue in that a number of drafts that had been requested to be adopted as WG drafts. Alex stated that advanced work could progress only after the work on the basic solution had been completed. Stewart Bryant disagreed with putting advanced work on hold pending completion of basic. SIDR - There are two areas which are being discussed. One is work on certificate profiles and repositories. The other is work to address deficiencies in TCP MD5: several drafts to allow key rollover and stronger authentication were discussed. A proper home WG for the TCP MD5 work needs to be found. The next work item for SIDR is BGP architecture: allow certificates to be signaled. VRRP - Not meeting at this IETF. VRRP for v6 is being reviewed by the IESG; some feedback from SEND has been requested with respect to the security section. The unified MIB is under MIB doctor review. The subsecond timer work is under WG review. 1264 obsolete discussion ------------------------ What should the requirement be for routing area documents? There was strong consensus to have "good" requirements. In the process of defining "good" it became clear that there is a general consensus that there is really no need to have special requirements for routing which are different from the rest of the IETF. At this stage it is proposed that it is a WG matter to determine requirements relating to implementation reports, and due attention should be given to quality of WG documents in this process. This would make RFC 1264 historic (not "obsolete"). IP routing in the GIG --------------------- Dow Street gave an extensive presentation on the USDoD "Global Information Grid" initiative. Initially, the engineering activity is mostly within the department of defense, but the expectation is that the scope will be extended to other military partners and government networks. This network is expected to push the limits of what current protocols can achieve in multiple dimensions, including routing, QoS and security. DoD is looking at this from a long term perspective with a goal of 100,000 routers in 12 years. There are at least two fundamental differences that distinguish GIG from the current Internet. One fundamental difference is the proposed pervasive node and network mobility: routers are on vehicles. This will have profound implications for inter-domain routing. The presentation listed many underlying assumptions in BGP which no longer hold in such an environment. Some examples include the fact that routers frequently move from one AS to another AS, the assumption that it possible to make a distinction between border routers and interior routers, the idea that it makes sense to define policies at the AS level, etc. Another fundamental difference which was discussed was the "routing commons", i.e. the cooperation model between the network domains. The GIG will be "mission-oriented": there will be much more emphasis on a single overriding authority which can allocate scarce resources to achieve a common goal. This was contrasted with the Internet which was suggested to use an "economic" model. Russ White remarked that there is probably more cooperation in the Internet than most people think. The GIG will be a long term project. The short term goal is to extend existing protocols including BGP. In the long term more fundamental architectural changes may be required such as a combination of IDR and MANET. The IETF was invited to contribute; this network will get built and will use commercially available products. IAB Workshop ------------ Dave Meyer reported on the proposed IAB workshop on routing. Mid-October is the likely date. Current workshop activity appears to be the definition of a routing problem statement and a requirement list. This is understood to be a by-invitation workshop. Ross Callon commented that it would be valuable for a broader discussion and call for input on the routing-discuss mailer on the identification of the problems that need to be solved wrt routing and addressing. The routing-discuss mailing list is routing-discussion at ietf.org List-Subscribe: , Yakov noted that the IAB had held a routing workshop some years back, and inquired about the differences between the previous IAB exercise and this one. Some hope was expressed that this would produce a productive outcome. Open Mike --------- No comments.