Administrivia
10min, Chairs

draft-anshuverma-bess-vpls-best-site-id-02
10min, Anshu (14:12)
Jorge: solution well documented. but all problems solved in vpls multihoming draft or do I miss something?
Anshu: usefull if you don't have the implem of vpls mh
Jorge: would you agree that if one has the vpls mh implem, this is not needed?
Anshu: yes
Ali: what precisely does help convergence ?
Anshu: if I bring in a pw
Wim: vpls mh is WG draft, so what would be the need for that? If need a solution in this space, implement vpls mh. Otherwise twice the interop efforts.
Jorge: we would burn a bit, maybe standardize this as Informational
Wen Lin: agree with Jorge and Wim
Patrice: if you care about convergence you should consider multi-homing instead of single homing

draft-ietf-bess-mvpn-fast-failover-01
5min, Robert (14:25)

Greg Mirksy: what do you mean by "BFD session is brought down", also multipoint BFD does not have a bootstrap
Rob: BGP routes are the bootstrap, just use admindown
Greg: need to bring this to BFD WG
(Wim, off mike): just stop sending keepalives
Rob: we'll clarify this part of the procedures
[discussion on BFD boostraping]
Greg: the work on bootstrapping p2mp BFD session belongs in BFD WG

draft-zzhang-bess-mvpn-evpn-cmcast-enhancements-00
20min, Jeffrey (14:34)

Ali, about slide 13 ("Segmentation w/ Untargeted Explicit-tracking C-multicast Route: Inter-as Example"): why not use a SMET route with an extended community, rather than a new route ?
Jeffrey: we thought about it, but it's not good enough, ok to discuss it offline
Ali: I agree that the procedures are useful, but we can leverage existing routes

draft-sajassi-bess-evpn-igmp-mld-proxy-00
10min, Ali (14:57)

Jorge: should use these procedures for single active mh for convergence
Ali: agree

draft-ietf-bess-l2vpn-yang-00
draft-ietf-bess-evpn-yang-01
draft-dhjain-bess-bgp-l3vpn-yang-01
10min, Patrice (15:08)

Thomas: fxc is not yet adopted, we would like you to consider an augmentation of the model rather to fit everything upfront
Patrice: ok
Ali: if straightforward to incorporate then it should
Thomas: we'll consider that
Iftekar: service tunnel notion should exist in the existing model, kind of a pw concept
Patrice: we should check current yang proposal if it fits for evpn vpws
Thomas: on bgp, might need to work with IDR
Thomas: on l3vpn yang, as said we are considering to call it for adoption
Patrice: ideas on how to progress Yang models faster would be helpful
Iftekar: would be good to have examples, use-case configuration examples, to show how it will work
Thomas: continue the discussion on the list
Iftekar: example on how this is used would be useful

draft-ietf-bess-dci-evpn-overlay-03
5min, Jorge (15:18)

Lucy: we are proposing an alternative solution
Ali: are you proposing to not look at the mac or ip but only the label and do stitching? That's inter-as option b.
Patrice: make sure to incorporate latest discussion points
Jorge: yes


draft-sajassi-bess-evpn-vpws-fxc-00
10min, Ali (15:24)

Iftekar: what about single port use-cases?
Ali: default case you need to support anyway
Thomas: who has read?
(6 hands)

draft-heitz-bess-evpn-option-b-00
10min, Jakob (15:35)

Tony: who is adding the PEO? ingress PE or ASBR ?
Jakob: the first router to change the nexthop adds the PEO ; if the ingress PE were to add the PEO, it would also work
Thomas: who has read?
(12 hands)

draft-snr-bess-pbb-evpn-isid-cmacflush-00
10min, Jorge (15:41)

Ali: cmac flushing on per isid and bmac. why not use existing mechanisms with an RT autoderived from the ISID ?
Jorge: still need an indication of isid in route
Ali: I agree with motivation, not the solution ; the meaning would be "if you send with the RT of the ISID, then flush"
Jorge: if you don't include isid, you'll flush for other isid
Thomas: please continue discussion on the list

draft-gmsm-bess-evpn-bfd-00
10min, Greg (15:50)

Jorge: I don't care which working group it should go, but BESS should be made aware anyway of that work. Also is there a mistake on the EL place in one of the slides
Wim: I would like to see bgp used to bootstrap bfd session, to fulfil the promise of evpn
Wim: discuss with BFD if bgp solution viable
Wim: also 2 solutions for data plane described, single one is preferred
Greg: of course
Rashed, BFD co-chair: 2 proposals one with IP header the other without, needs to be clarified
Thomas: WG and chairs of BFD & BESS will need to talk together