IETF 93: Open Shortest Path First (OSPF) WG Agenda Wednesday, July 22nd, 2015. 17:40 - 19:40 Afternoon Session III Location: Karlin I/II ======================================================== Chairs: Acee Lindem Abhay Roy Scribe: Les Ginsberg (ginsberg@cisco.com) WG Status Web Page: http://tools.ietf.org/wg/ospf/ 1) Administrivia - 5 minutes - Blue sheets - Scribe/jabber - Jabber room: ospf@jabber.ietf.org 2) WG Status Update - 15 minutes See slides draft-ietf-ospf-routable-ip-address Acee: Do we still need this? Uma Chunduri: Pre-dates prefix/link attributes draft. Prefix-link attributes is better solution draft-ietf-ospf-sbfd-discriminator-01 Les: I think S-BFD has gone through last call - so you can progress OSPF S-BFD (Post meeting update: Base S-BFD spec is currently in last call: http://datatracker.ietf.org/doc/draft-ietf-bfd-seamless-base/ 3) H-bit Support for OSPFv2 - 10 Minutes - https://datatracker.ietf.org/doc/draft-keyupate-ospf-ospfv2-hbit/ - Padma Pillay-Esnault See slides Chris Hopps: Similar to OL bit in IS-IS. Sometimes there are reasons to advertise non-local addreesses. Acee: How should you handle non-local routes? Stub routers treat them as local. Could use R bit. Unlikely to have routing loop - so this closes that unlikely case. Padma: R bit is reverse of this and we may want to relax the backward compatibility requirements in some deployments. Acee: Ok Les: Why do you specify the router "must not act as forwarder"? Might as well try to forward traffic which arrives. Padma: This can be changed. 4) OSPFv2/v3 SR Update - 10 minutes - https://datatracker.ietf.org/doc/draft-ietf-ospf-segment-routing-extensions/ - https://datatracker.ietf.org/doc/draft-ietf-ospf-ospfv3-segment-routing-extensions/ - Peter Psenak See slides Uma: Want to see MPLS dataplane and IPv6 dataplane for OSPFv3 Peter Psenak: Need to wait for overlap SRGB conflict reslution decision before going to WG LC. 5) OSPF BIER Extensions (BIER WG Draft) - 10 minutes - https://datatracker.ietf.org/doc/draft-ietf-bier-ospf-bier-extensions/ - Peter Psenak See slides Acee: BIER WG owns this document. 6) OSPF Link Overload - 5 Minutes - https://datatracker.ietf.org/doc/draft-hegde-ospf-link-overload/ - Shraddha Hegde See slides Peter: Consider LLS. Advertisement should be linked scope only. Chris Hopps: Big metric will be advertised - isn't that enough to support TE C-SPF? Shraddha: TE uses other link attributes such as bandwidth. Chris Fowler: Consider a solution which would also lower the advertised available bandwidth (for example). Les: Proposed solution requires passing the OL signal to local TE as well - not just controllers. Chris Hopps: Is it always the case that the constraints need to be signaled indirectly? Les: You could do the same thing for cost to neighbors by not changing metric - instead advertising a new state flag - but this would be very invasive/not backwards compatible. Treating TE attributes differently than metric is an inconsistent model. 7) OSPF Self-Defined TLVs - 10 Minutes - https://datatracker.ietf.org/doc/draft-chunduri-ospf-self-defined-sub-tlvs/ - Uma Chunduri See slides Acee: Make it clear that you are depending on administrator configuration Acee: Some use cases sound like they should be stanadardized (e.g., standard telemetry data). Les: Using examples that have nothing to do w routing does not make me want to support it. Chairs: Similar to IS-IS GENAPP Les: GENAPP should be used in separate protocol instance. Uma: SFC does not want standardized codepoints. Acee: To the room: "If we took this up who would implement it?" Withdraw the question as premature. 8) Advertising Tunneling Capability - 10 Minutes - https://datatracker.ietf.org/doc/draft-xu-ospf-encapsulation-cap/ - Xiaohu Xu See slides Ahmed Bashandy: Want to advertise capability of being the tail end of a tunnel. Decap capability is better terminology. Why not use admin tags? Xiaohy: Will consider this Acee: Who thinks it is useful? (Few hands) Acee: Discuss on list. 9) OSPF Yang Data Model Update - 20 Minutes - https://datatracker.ietf.org/doc/draft-ietf-ospf-yang/ - Derek Yeung See slides Jeff Tantsura: Draft­litkowski­spring­sr­yang­01 has been accepted as WG document - please update name. Uma: Where to put prefix/ADJ Sids? Needs to be in seperate group. Derek: We are just reusing grouping definition (template) defined by YANG model Acee: Decided SR belongs in separate model based on deployments. Acee: Weekly call has good attendance - but could use more attendance. There are implementations being developed. 10) OSPF TTZ Update - 5 Minutes - https://datatracker.ietf.org/doc/draft-ietf-ospf-yang/ - Huaimo Chen See slides Acee: This is experimental track. When stable will assign some reviewers. 11) Extensions to OSPF for Temporal LSA - 10 Minutes - https://datatracker.ietf.org/doc/draft-chen-ospf-tts/ - Huaimo Chen See slides Acee: You are assuming that past link utilization is an indication of future reservations? Huaimo: Yes Acee: Like TDM :-) Chris Hopps: Do you have a description of how CSPF should use this info? Should this be tied into RSVP? Huaimo: Yes Chris Hopps: If there is not enough bandwidth in an interval a proposed constrained path would be rejected? Huaimo: Yes - can reserve bandwidth at certain times of day. Acee: TEAS conflicted w RTGAREA- therefore we did not see base architecture presented there. Wait to see how architecture progresses. http://tools.ietf.org/html/draft-chen-teas-rsvp-tts 12) OSPF TE Values for non-TE Applications Alternatives - 10 Minutes - Peter/Acee See slides Acee: SRLG requires multiple LSAs. For TI-LFA this seems a better encoding for non-TE cases. Chris Hopps: How would you define what SRLGs you advertise? Peter: Only support local SRLG now. Shraddha: If TE is enabled as well then info will be advertised twice? Peter; Yes Shraddha: May waste space. Peter: Yes - takes up more space in such cases. Shradda: Do I have to look for duplicate info? Acee: Yes - if you are using both TE and non-TE applications. Jeff Tantsura: Support this work - this is a longstanding problem. Abhay: Out of time - take to list