MPLS WG minutes (session 1 - Wednesday, July 22, 2015 - 1740-1940): MPLS YANG activities report (Tarek Saad) Showing a compiled report on all MPLS related YANG models active & expired and overall status on each. Highlighting any areas that have not been touched on, and give advice. Questions / comments: - Eric Grey: what is meant by outstanding work: Tarek: no I-D active draft with MPLS WG - David Sinicrope: recommendation to authors to add intention to create YANG I-Ds to wiki in section for NEW YANG models - Stephane Litkowski: with regards to attachment to mpls in Xpath, we are working at the RTG WG to define a meta model, you should see where your MPLS model fits in. - David Sinicrope: for models that have expired, ask authors to mark as abandoned - Loa: is there any areas (as a WG) that are missing not covered? Tarek: besides what was noted, not aware of any MPLS LSP YANG model - draft-zheng-mpls-lsp-ping-yang-cfg (Greg Mirsky): Jeff Haas: divide work/model into core/basic features and optional ones to be protected by "if-feature" Loa: - Poll for how many? read/support adopting I-D > 6. a bit weak, need to socialize more on the maling list Discuss with David Sinicrope on the review expected around the time of WG adoption - David Sinicrope: - WG is expected to own the model, authors/chairs to decide what other WG is expected to review the model - Loa: have number of drafts exist that I want MPLS WG to be consulted - David: One WG will take the lead and others to review - Loa: need way to allow WG to identify drafts (from other WGs) to be reviewed by the WG - Jeff: there is a tool that may help in this - Himanshu: overal between i2rs and other WG work - Loa: to defer this discussion - David: Loa, George: To be clear I wasn't saying the lead had to populate the Relevant WG column of the table. Specifically the instructions on the wiki state " For the "Relevant WGs" column where the drafts should be reviewed - add the WG abbreviation for each WG where the draft should be reviewed. (i.e., where the WG has relevance to the model/draft) Please bold the primary WG responsible for the model. Also change Status to "Not Reviewed". If you have questions on which WGs should review or which should be designated as primary, please discuss with relevant WG Chairs." just added "NOTE: WG Chairs - if you think your WG should be listed as relevant please add it after discussing with the authors." to the instructions to address your point in the meeting. YANG data model for LDP and mLDP - draft-raza-mpls-ldp-mldp-yang (Xufeng): Robin: mldp for mvpn services, do you expect model to have tunnels?.. to continue the discussion with authors offline.. Loa: Ina do you have update on openconfig LDP model? Ina: concentrated so far of RSVP-TE model, but for LDP expect ietf model to adapt to add support for operational state for each config node (config/state leaves) Xufeng: we plan to add support for config state, but ongoing discussion on where to add it in model Ina: if the model does not adapt config/state nodes at the bottom containers, then open-config model can not import from it Discussion on YANG activities (WG/chairs): Himanshu: there is overlap of i2rs work and other WGs (e.g. MPLS), e.g. TE topology, do we have a process? Lou: the meeting that happened with i2rs was narrow scope, limited to TE topology -- up to both sides to sort it out how to proceed Loa: asking for more structured approach/process Jeff: as engineers let's work it out... chairs expected to facilitate the discussions Jeff: do the work (don;t optimize initially), duplicate data may exist, identify and work it out between parties. Who owns this stuff longterm? Base models and extensions Loa: instruct teams working on YANG models to reach out and see if there is overlap George: ask authors a) to make the chairs aware and b) work with other party to resolve the overlap David: if use the wiki (initent to create I-Ds), then identifying overlap is easier George to David: has the issue of 2 groups viewing the data in two different ways come up? draft-esale-mpls-ldp-node-frr (Santosh Esale): Robin: interesting solution, need to use RSVP and LDP together for this Stewart: what's new in this? what about loop prevention Santosh: procedures introduced here let you to restore the NH and MP automatically. Micro-loops are still possible Loa: continue discussion on the list draft-chandra-mpls-ri-rsvp-frr (Chandra Ramachandran): Loa: How many read the draft? A: good number Loa: Chairs will take a look at this and George: this has co-dependency on another draft draft-bryant-mpls-synonymous-flow-labels (Stewart Bryant): Loa: on the requirement draft, and let it die I'd rather make it WG document and let it die correct or complete, more concerned about correct than complete Luang: comment on ECMP, presented 2 options there's is a 3rd option, use IP hash George: the payload may not be IP Stewart: control system how much work to do before asking for WG adoption Loa: poll on how many read the draft - how many think it deserves to be adopted MPLS WG minutes (session 1 - Thursday, July 23, 2015 - 1520-1720): draft-mtaillon-mpls-summary-frr-rsvpte (Tarek Saad): Robin: why not use RSVP Hellos to indicate bypass active Tarek: RSVP Hellos are not necessarily enabled always. Hello session between PLR/MP does not necessarily indicate that all bypasses (between PLR and MP) are active Loa: poll on how many read the draft and willing to adopt draft-fang-mpls-hsdn-for-hsdc (Luyuan Fang): Questions / comments: Luang: ask for WG adoption Loa: poll on how many read the draft and support WG adoption Stewart: what are advantages of this over SR? Luang: this is new architecture George: please follow on this on the list draft-fang-mpls-label-forwarding-no-swap-00 (Luyuan Fang) Questions / comments: George: this is internal optimization Loa: to take this discussion to the list draft-decraene-idr-next-hop-capability-01 (Bruno Decraene): Questions / comments: Robin: useful, but propose new challenges (need to exchange other forwarding capabilities) Kireeti: the goal is not to advertise every node capability with BGP Loa: do you plan to continue working on this in IDR WG or bring it to MPLS? Loa: send a mail to list to explain why this should or not be continued in MPLS draft-ietf-mpls-lsp-ping-mpls-tp-oam-conf (Greg Mirsky): Questions / comments: Greg: document is ready for WG LC George: you can request WG LC draft-mirsky-mpls-residence-time (Greg Mirsky): Questions / comments: Loa: poll for how many read and support WG adoption Loa: please send mail asking for WG adoption draft-ietf-mpls-self-ping (Ron Bonica): Questions / comments: Ron: forward to IESG Loa: discussion that this has progressed too fast Ron: we had a lot of comments that were addressed Loa: will go ahead, expect in 1 or 2 weeks something will happen draft-farrelll-mpls-opportunistic-encrypt (Adrian Farrel): Stewart: does this actually improve security? Adrian: if using Diffie–Hellman, removes one party from the key exchange so potentially more secure michel: what is missing is state of the art of existing mechanisms Adrian: there's a place holder in the document, if someone wants to describe it we can add in document draft-cheng-mpls-tp-shared-ring-protection (Liang Geng): Liang: would like to ask for WG adoption Loa: poll on how many read and support the draft for WG adoption (fairly good number) Loa: please notify when you have the updated version draft-kompella-mpls-rmr-01 (Kireeti Kompella): Kireeti: draft is ready for WG adoption Loa: when you're done with new version notify and we'll proceed with WG adoption draft-kompella-mpls-larp-03 (Kireeti Kompella): Kireeti: with reviewers sign-off we will republish and ask WG for adoption draft-kompella-mpls-rsvp-ecmp (Kireeti Kompella): Kireeti: change to ECMP to do unequal cost multipath Ravi: why do we need association object for sub-lsps, need new association type c-type(?) Lou: association type of C-type?