Minutes IETF105: grow
minutes-105-grow-00

  IETF 105 GROW meeting agenda

Administrativa
==============

    - assign jabber scribe (TBD)
    - assign minute taker - Jared

    Current documents:
        * bmp-out is in IESG review, authors working address comments
        * draft-ietf-grow-wkc-behavior is with the RFC editor
        * draft-ietf-grow-bmp-local-rib waiting for concensus in WG (Jeff?)
        * WGLC for draft-ietf-grow-bmp-registries-change (positive support)
        * draft-scudder-grow-bmp-peer-up was accepted as WG document

    New Documents:
        * call for adoption of draft-sa-grow-maxprefix

    Other News:
        * IDR is looking to re-charter and define milestones, GROW participants
          should think about what they’d want from IDR in the next 5 years
        * Living documents update (renamed “Evolving Documents”)

Agenda
======

Alexander Azimov (15 minutes):
        - leak-detection-mitigation

Paolo Lucente (10 minutes):
        - draft-ietf-grow-bmp-local-rib
        - draft-lucente-grow-bmp-tlv

Camilo Cardona (10 minutes):
        - draft-cppy-grow-bmp-path-marking-tlv-00

Guyunan (10 minutes):
        - draft-xu-grow-bmp-route-policy-attr-trace
        - draft-gu-grow-bmp-route-leak-detection

AOB
===
    5 minutes

— snip —

Job: IDR is rechartering, please be aware of it
John Scudder: We welcome feedback at our Meeting on Friday for IDR rechartering

Q (via jabber) - What IDR documents go to which WG
Scudder: That was a topic this morning with the AD, please bring them to the

Job: There will be an update on Living Documents tomorrow at the Plenary
(Warren: No) Job: They will be meeting in the same room tomorrow about Living
Documents 830a

- Presenter Alex:

Job: I recommend you put TBD vs values that have not assigned
Jeff Haas: Large communities were not intended to have well known values, think
about if it should be in rfc1997 space as there are IANA considerations and IDR
consultation should occur

Doug Montgomery: I think what we agreed was the default policy would be MUST be
reject Alex: You may or not want to mitigate Doug: Lets try to be more clear
about it

Ben Maddison: I would not dress up mitigation, we should do this by default

Loc Rib
Presenter: Paolo

No questions

draft-lucente-grow-bmp-tlv
Presenter: Paolo

No questions

draft-cppy-grow-bmp-path-marking-tlv

Presenter: Camilo

Jared: The types (bitfield) where should this set of types live, in IDR as a
protocol thing or elsewhere?

Jeff Haas: The problem with a lot of things with this tracing the BGP pipeline
are not fully synchronous, like reachability or next-hop may come very late in
the process.  If it’s selected for best-path may be much further down the
pipeline.  Trying to tie these things in, locks the implementation down to send
things out much later Camilo: Loc-rib is better? Jeff Haas: local-rib is far
worse as you may have endless churn.  There is something interesting here, but
the primary feedback is about putting it into BMP, it’s possibly choking the
firehose.  Perhaps this is best served as a telemetry feed.

Ben Madison: At least one implementation of add-path we run has a way to mark
routes that are part of a set.  This fits in well with the set-types you
describe here.

Alex Asimov: We have many other things like ASPA that are impacting the
implementation.  We may have reasons that may reject due to route-leak, RPKI
failure Camilo: I don’t see why not. Haas: Maybe it’s easy, but it’s expensive

Chair: We are time constrained, can we take to the list

draft-gu-grow-bmp-route-leak-detection
Presenter: Yunan Gu

BMP for BGP route leak detection

Not much has changed since IETF 103 presentation. Detection of route-leak has
happened with the local as.

No questions

draft-xu-grow-bmp-route-policy-attr-trace
Presenter: Yunan Gu

Chair: Please send feedback to the list.