ISIS WG, 1510-1610, Garden Ballroom 3 Chairs: Dave Ward, Chris Hopps Notes: Donald Eastlake Blue sheets passed around. Agenda & Document Status, Dave Ward ----------------------------------- There were no requested changes or additions to the agenda. WG LCs: ISIS Layer2 suite: draft-ietf-isis-layer2-06.txt, new draft posted with corrections. draft-ietf-isis-trill-01.txt, See AD proposal. draft-ietf-isis-ieee-aq-00.txt, see interop report this meeting. IS-IS Purge TLV - Passed draft-ietf-isis-purge-tlv-04.txt IS-IS Registry Extension for Purges - Passed draft-ietf-isis-reg-purge AD Followup: draft-ietf-isis-bfd-tlv-02 IESG Review - revised ID needed: draft-ietf-isis-genapp-03 IESG Approved: draft-ietf-isis-ipv6-te-08 New drafts - not WG draft yet: draft-nie-isis-mi-mib-00 draft-shen-isis-oper-enhance-00 draft-amante-isis-reverse-metric-00 Errata cleared up: RFC 5303 Reclassification proposal RFC 1142 to Historic Progressing draft-ietf-isis-trill --------------------------------- Stewart Bryant, Ralph Droms A new document draft-ietf-trill-adj will be added as a normative reference in draft-ietf-isis-trill. draft-ietf-isis-trill will then proceed to the IESG for review with the normal IETF process. The "adjacency document" will have state machine description of Adjacency formation DRB (aka DIS) election Rules for two-way and MTU matching for advertisements Creation and use of pseudo-nodes NOTE: Document to be in TRILL WG. TRILL WG Chair agreed that ISIS will get notice of draft submission/iteration for easier review and communication. Stewart Bryant - Given that the TRILL specification is in good shape, the new document should be something that any junior programmer could do. Dave Ward - What is important is through tehchnical review. Hopefully the above items will cover the major points. Stewart Bryant - No, the TRILL document is what it is. What this will provide is clear documentation. If a serious technical problem is found, it needs to be fixed. But complaints about the TRILL document being the wrong color or the like are not relevant. Unless there is a serious techncial problem, the documents should move through swiftly. Brief Update on First SPBM Interop, Paul Unbehagen -------------------------------------------------- Issues Encountered - minor. Review of test configuration. 2 equal cost paths. Donald Eastlake: There is a White Paper on the TRILL interoperability test at the University of New Hampshire Interoperability Laboratory (UNH IOL) linked off the www.postel.org/rbridge page (which is in turn linked from the TRILL WG home page). Dave Ward: Please post it to the list. IS-IS Operational Enhancements for Network Maintenance Events ------------------------------------------------------------- draft-shen-isis-oper-enhance-00, Shane Amante Problems to address: Requirement to gracefully shutdown the IS-IS protocol on a LAN interface. Bi-directionally remove one or all nodes attached to a multi-access LAN from the ISIS topology. Next Steps Dave Ward: Anyone think this should not be a WG draft? (No response.) We will take it to the list for confirmation but it looks good. Reverse Metric TLV, Shane Amante -------------------------------- draft-amanate-isis-reverse-metric-01 Way to direct one or more neighbors on a point-to-point or multi-access link to increase the metric they report for a link. Dave Ward: Anyone think this should not be a WG draft? (No response.) Let the record show there was no opposition to either draft. Question: ~ what about router that send out bad TLVs of this type? ~ Shane Amante: If you have a bad actor, you have big problems. You should use authentication in a production network. If someone steals the passwords, you have bigger problems than this TLV. Question: Have you thought about what happens if the DIS changes from one aware of this TLV to one not aware of it. Amante: Yes, we have thought about that. There are solutions. As an operator you need to be aware of this situation and insure you have priorities set so the back-up DIS will be able to do the right thing. Tony Li: Problem... some vendors have said no. So you really want all systems in your network upgraded to support this before you turn it on. Amante: Yes, that's a consideration. Dave Ward: Should something be added to the draft about this. Amante: Not clear that we need more complexity to handle this. The operator should be aware what's happening in their network.