=============================================================================== PCE Working Group Meeting IETF 98 (Chicago, IL) Working Group Chairs: Julien Meuric (julien.meuric@orange.com) JP Vasseur (jpv@cisco.com) Jonathan Hardwick (jonathan.hardwick@metaswitch.com) Working Group Secretary: Daniel King (daniel@olddog.co.uk) Responsible AD: Deborah Brungard (db3546@att.com) =============================================================================== Session I Time: March 27, 2017, 15:20-16:50 (3:20pm-4:50pm) Location: Zurich G, Swissotel, Chicago IL, USA With thanks to all our scribes! ------------------------------------------------------------------------------- 1. Introduction --------------- 1.1. Administrivia, Agenda Bashing (chairs, 5 min) - The session is being co-chaired by Jon Hardwick and Julien Meuric. Unfortunately, JP was unable to travel to the IETF this time. - Agenda bashing: no agenda changes. 1.2. WG Status (chairs, 15 min) Jon Hardwick: Stateful PCE draft is with the IESG and pending an update; the authors have promised an update by 1 April (no joke). Jon Hardwick: LSP initiation draft is with the IESG and is pending routing directorate review before IETF last call can be issued. Jon Hardwick: Inter-area/AS applicability draft needs an update before we can submit it to the IESG. Authors are working on it. Jon Hardwick: PCEPS draft is now being shepherded by Cyril Margaria; he has posted some comments and needs to see a new revision before we can submit to othe IESG. Jon Hardwick: WSON-RWA draft has been shepherded by Daniele Ceccarelli (thanks Daniele!) and is now blocked on advancing by the base PCEP-GMPLS draft. Jon Hardwick: Base PCEP-GMPLS draft is pending a shepherd report from Julien. Jeff Tantsura: Segment Routing draft will be updated this week. Jon Hardwick: IANA early allocation for this and LSP-setup type drafts expires in August - we must publish before then. Dhruv Dhody: Stateful P2MP draft can be progressed after the 6006bis draft. Dhruv Dhody: We are discussing with association-group authors, to make sure it could be prioritized, as multiple documents depend on it. Young Lee: The Stateful PCE GMPLS draft will be revived; Xian has moved on but she will hand the pen to someone else. 2. Work in Progress ------------------- 2.1. PCEP Extension for LSP Diversity (Stephane Litkowski, 5 min) https://datatracker.ietf.org/doc/draft-ietf-pce-association-diversity/ No questions. 2.2. Association-aware Computation (Stephane Litkowski, 10 min) https://datatracker.ietf.org/doc/draft-litkowski-pce-state-sync/ Andrew Dolganow: Why assume that the last state received is the final state? Message routes through the network may be different. If sessions are flaping then it may get out of sync. Stephane Litkowski: We had an idea to add a timestamp. Robin Li: In multiple failure scenarios, this may get out of sync. If the session between PCC and PCE is broken or if PCC is restarted. The impact on other protocols like netconf, BGP needs to be analyzed. Stephane Litkowski: You can not solve all the problems. Julien Meuric: Who has read the document? [Not many except authors] Julien Meuric: Please read document and respond to the list. 2.3. PCE's Control Request (Chaitanya Yadlapalli, 5 min) https://datatracker.ietf.org/doc/draft-raghu-pce-lsp-control-request/ Julien Meuric: Please do not suggest a specific bit number. This is the role of IANA. Just use TBD in your draft. Andrew Dolganow: Can you clarify the need for this, before you ask for a bit. Chaitanya Yadlapalli: We will try to clarify the use case in the I-D. Julien Meuric: Please update the document and reflect changes to the list. Julien Meuric: Who has read the document? [About 12-15] Who, of those that read the document, think this is a useful feature? [About the same] 2.4. Experimental Codepoints (Dhruv Dhody, 5 min) https://datatracker.ietf.org/doc/draft-dhody-pce-pcep-exp-codepoints/ Andrew Dolganow: Good work. There will be expected behavior, can you describe how to interpret the experimental bits? Dhruv Dhody: Actually, I think this is already discussed in the PCEP RFC [RFC5440] Andrew Dolganow: Can you just clarify this. Dhruv Dhody: Yup Jon Hardwick: Poll, who read the I-D? [8] Who thinks we need this? [same number] 2.5. Stateful H-PCE and ACTN (Dhurv Dhody, 10 min) https://datatracker.ietf.org/doc/draft-dhody-pce-applicability-actn/ https://datatracker.ietf.org/doc/draft-dhodylee-pce-stateful-hpce/ Michael Scharf: [Slide 9] You extend PCEP (on MMI and MPI interface) for the VN concept, but it's not clear on the architecture. This is an ACTN issue, not directly related to the PCEP WG or PCEP I-D. Dhruv Dhody: We will pick this up in the ACTN architecture I-Ds Jon Hardwick: Who read the ACTN applicability document [10] Who thinks we need this? [About the same] Jon Hardwick: Who has read the stateful H-PCE document [About 15] Who thinks we need this? [About the same, a few less] Jon Hardwick: We will confirm on the list. 3. New I-Ds ----------- 3.1. PCE-initiated BRPC LSP Setup (Julien Meuric, 10 min) https://datatracker.ietf.org/doc/draft-dugeon-brpc-stateful/ Fatai Zhang: Interesting I-D. Are you trying to use PCEP to create end-to-end connections across domains, i.e., like RSVP-TE. Julien Meuric: Yes. Fatai Zhang: Can we use these PCEP extensions for other applications? Julien Meuric: Yes, segment routing is a use case. Fatai Zhang: Could you use PCECC? Julien Meuric: I think there is some common ground. We should discuss. Dhruv Dhody: I like the I-D, I like the SL-label to be generic, so that it can be useful for stateful H-PCE as well, and a PCECC-like approach might be better. Dhruv Dhody: We need to investigate the security aspects of stateful BRPC, where PCE relationship is on the peer level, not client-server. Julien Meuric: Agree. Its a -00 version. 3.2. PCEP Extension for Associated Bidirectional LSPs (Rakesh Gandhi, 10 min) https://datatracker.ietf.org/doc/draft-barth-pce-association-bidir/ Jon Hardwick: Who thinks this is a good I-D to have? [12] Good, we will need to see more dicussion on the list. 3.3. PCEP Extensions for PM Reporting (Rakesh Gandhi, 10 min) https://datatracker.ietf.org/doc/draft-gandhi-pce-pm/ Andrew Dolganow: Capability is useful, but adding telemetry data to PCEP will not scale. The PCE is responsible for path computation which is already compute intensive. You can't add real-time telemetry without overloading it. Rakesh Ghandi: Motivation is to use the same channel to get LSP state as well as performance. Fatai Zhang: Relationship to PCEP-LS is not mentioned in the document. Jeff Tantsura: Same concerns as Andrew, do you measure at head-end or per node. Rakesh Ghandi: This is for head-end PCC, and for the LSPs that PCE cares about. 4. I-Ds Pending Further Discussion 4.1. Bandwidth Scheduling (Yan Zhuang, 5 min) https://datatracker.ietf.org/doc/draft-zhuang-pce-stateful-pce-lsp-scheduling/ Yan Zhuang: The authors have now produced a merged document and are requesting WG adoption. Julien Meuric: New version was only uploaded today, so we must give people time to read and comment before polling for adoption. Please send comments to the list. ------------------------------------------------------------------------------- End of session 16:50.