Routing Area Director: o Joel Halpern: jhalpern@newbridge.com Area Summary reported by Joel Halpern/Newbridge Networks Corporation Inter-Domain Multicast Routing Working Group (IDMR) The IDMR Working Group met over two sessions in Stockholm. The presentations fell into two categories: those concerning ongoing implementation and simulation work, and those regarding hierarchical multicast. It was evident that implementation and simulation work both in CBT and PIM has been the subject of much work. Bay Networks is currently implementing CBT and hopes to release it this summer, whilst Cisco continue with their PIM implementation. Workstation implementations continue in both PIM and CBT. Simulation results conducted by Harris Corporation once again showed that shared trees compare favourably (especially in terms of delay and complexity) with shortest-path trees. Original estimates of very poor delay characteristics of shared trees have proved inconclusive. The recent growth of the MBone (its size has doubled over the past 8 months) has meant that it was necessary to focus on the architectural and engineering aspects of the MBone so that it can survive this growth. Deering et al. recently proposed Hierarchical Multicast whereby the MBone should be divided into non-overlapping `regions', with Level 1 routers operating one multicast routing protocol inside a region (intra-region), and Level 2 routers operating DVMRP between regions (inter-region). The presence of DVMRP at Level 2 is considered an incremental first step in the deployment of hierarchical multicast for the short term. The proposal for its implementation at Level 2 in the short term is due to the extensive implementation experience we have had with DVMRP. However, Level 2 DVMRP should eventually be replaced by whichever IDMR protocol is chosen by the working group as being the most effective in terms of scalability, simplicity, and the (non-)complexity of its interface with Level 1 routers. Hierarchical CBT was presented, showing how CBT can interoperate with a variety of multicast protocols both at Level 1 and Level 2. A similar presentation was made for PIM. It was especially evident from these presentations that the issue of core (or RP) advertisement and selection needs further investigating with the goal of finding a scalable solution. The consensus of opinion was that much more implementation experience is needed in order to establish a Level 2 protocol, and also to establish whether multiple Level 1 protocols can co-exist. This is very much dependent on whether a simple, uniform, and non-complex Level 1/Level 2 interface can be defined. Inter-Domain Routing Working Group (IDR) Agenda for the Tuesday IDR meeting: o Agenda bashing o BGP-4 to Standard o DPA proposal (Tony Bates, MCI) o Route Server to Experimental o Route Damping proposal Yakov will issue a last call to the group within a month for all the BGP-4 documents. The DPA proposal was reviewed, and two vendors have indicated their intent to implement this proposal. The Route Server document works on solving the IBGP full mesh problem. The document will move to Experimental, and has been implemented by one router vendor. An alternative solution to the IBGP full mesh problem has been implemented by Cisco systems, and will be written up prior to the next meeting. Route flapping has become an issue in the Internet. Curtis Villamizar wrote up a document on how to damp out route flapping two years ago. This algorithm has implemented, and an updated document will be forward to the group by next IETF. IP Routing for Wireless/Mobile Hosts Working Group (MOBILEIP) The Mobile IP Working Group met three times in Stockholm; once jointly with the IPng Working Group (IPNGWG) to discuss mobility for IPv6; and twice to continue to reach closure on mobility for IPv4. The next two paragraphs summarize each of these respective topics. IPv6 -- The consensus seemed to be that: (a) IPv6 mobility should exploit the enhanced functionality of IPv6 (as compared to IPv4) and should not be simply a transliteration of IPv4 mobility; and (b) that the MOBILEIP Working Group's charter should be modified (if necessary!) to include responsibility for IPv6 mobility. That is, mobility for IPv6 should be `owned' by the MOBILEIP Working Group. IPv4 -- Patent issues are still plaguing the MOBILEIP Working Group. At the request of the co-chairs, Joel Halpern (Routing Area Director) will request funds from ISOC to have legal counsel investigate the scope of infringement of the current working group draft on the patent by Charlie Perkins (IBM). Otherwise, the substantive content of the draft seems ready for implementation; in fact, many working group members plan to test inter-operability of their respective implementations, tentatively scheduled for the end of August. The group agreed that a wholesale rewrite of the draft (editorial changes not content changes) was in order. New Internet Routing and Addressing Architecture Working Group (NIMROD) The NIMROD Working Group met on 19 and 20 July. Most of the meeting time was taken with the following presentations: o Overview of the Implementation Model o Agent Discovery Protocol o Path Set-up Protocol o Reliable Transaction Protocol o Query Response and Update Protocols The main point of discussion was on connectivity restrictions between agents of a node and the agents of its subnodes. The consensus of the group was to that the `tight' restrictions were too confining and that a way to make a component node act as a `virtual' forwarding node for the parent node should be supported---for example, as part of the agent-discovery flooding protocol. Routing Over Large Clouds Working Group (ROLC) The ROLC Working Group met in Stockholm, with 87 attendees. The group substantially completed work on NHRP Version 1, which includes host-host, host-router, router-host, and limited router-router operation. Version 05 of the NHRP specification will be produced following the meeting, and after obtaining final consensus on the list, will be submitted for consideration as a Proposed Standard. Internet-Drafts discussing the full router-router case were solicited for the next IETF meeting. The group also discussed the NHRP Applicability Statement, Protocol Analysis, and MIB documents, and will continue work on them, to conclude at the next IETF meeting.