Open SPF-based IGP Working Group Chairpersons: Mike Petry/UMD and John Moy/Proteon CURRENT MEETING REPORT Reported by Rob Coltun AGENDA The OSPFIGP working group met for a half day on July 28th at Stanford. The agenda was as follows: o Implementations o Spec Changes o Net Management Items o What's Next ATTENDEES 1. Baker, Fred/baker@vitalink.com 2. Bierbaum, Neal/bierbaum@vitalink.com 3. Blackwood, Craig/craig@hprnd.rose.hp.com 4. Blumenthal, Steve/blumenthal@bbn.com 5. Coltun, Rob/rcoltun@trantor.umd.edu 6. Deboo, Farokh/fjd@bridge2.3com.com 7. Deering, Steve/deering@pescadero.stanford.edu 8. Doo, Way-Chi/wcd@bridge2.esd.3com.com 9. Farinacci, Dino/dino@bridge2.3com.com 10. Fuller, Vince/vaf@jessica.stanford.edu 11. Honig, Jeffrey/jch@sonne.tn.cornell.edu 12. Hytry, Tom/tlh@iwlcs.att.com 13. Ilnicki, Ski/ski 14. Jones, Bill/jones@nsipo.nasa.gov 15. Jordt, Dan/danj@cac.washington.edu 1 16. Karn, Phil/karn@thumper.bellcore.com 17. Medin, Milo/medin@nsipo.nasa.gov 18. Moy, John/jmoy@proteon.com 19. Oattes, Lee/oattes@utcs.utoronto.ca 20. Oran, David/oran@oran.dec.com 21. Petry, Mike/petry@trantor.umd.edu 22. Pugh, Rex/pugh@hprnd.rose.hp.com 23. Reilly, Michael/reilly@atari.nac.dec.com item Smith, Tom/toms@hprnd.rose.hp.com 24. St. Johns, Mike/stjohns@beast.ddn.mil 25. Stone, Geof/geof@network.com 26. Veach, Ross/rrv@seka.cso.uiuc.edu MINUTES 1. Milo Medin presented John Moy's slides of the Proteon OSPF implementation in John's stead. 2. Rob Coltun presented slides of UMD's BSD OSPF implementation. 3. During the presentations questions came up concerning the migration from RIP to OSPF. It was agreed that the best way to approach this is to initially put OSPF at the borders of the Autonomous System and work towards the center. We are considering writing a paper addressing this issue. 4. Questions also came up regarding running OSPF on BSD systems that currently do not support Multicast. We explained that the BSD version allows configuring non-broadcast, multi-access networks over broadcast networks to serve as a temporary fix for those systems that do not support Multicast; this option will remain in the implementation until Multicast is generally available. 5. There were discussions on conditions that may occur in very large Autonomous Systems such as how to handle routing updates when a gateway has run out of memory and on possible consequences of running the spf algorithm each time a new LSA arrives when a partition has been repaired. 6. We talked about some of the changes to the specification such as packet formats, the Fletcher checksum, hello packets on non-broadcast multi-access networks, and hello packets over point-to-point links. 2 7. We presented some possible MIB variables. These basically were state, database and packet summary information. Not much discussion. 8. Some discussion followed about the OSPF working group being close to (or at) the end. We talked about starting a Multicast routing working group.