Minutes IETF110: panrg
minutes-110-panrg-00

PANRG Agenda IETF 110
When: 15:30 - 16:30 UTC+1/14:30 - 15:30 UTC, Thu March 11th, 2021
Where: https://meetings.conf.meetecho.com/ietf110/?group=panrg&short=&item=1
Chairs: Jen Linkova and Brian Trammell
Minutes Taker: Jake Holland
Jabber Scribe: Brian Trammell
Jabber: panrg@jabber.ietf.org

Time (UTC+1) Length What Who
15:30 5m Welcome, Note Well, Agenda Chairs
15:35 5m Updates on draft-irtf-panrg-questions B.Trammell

    *   definition of path is a good question from list. discuss where it
    belongs after path properties.

15:40 5m Updates on draft-irtf-panrg-what-not-to-do S. Dawkins

    *    Brian, re “better way than just 1 chance?”: maybe “one chance per 2
    decades” *    Gorry: maybe “one chance per equipment lifetime”. Will read
    it and send proper comments *    Martin: is this actionable? *    Spencer:
    maybe something like broadening the tests you run, thinking outside the box
    *    Colin: +1 on “per equipment lifecycle”, and can it say something about
    structured testing? When deploying again, they did it in incremental way
    while working with the operators. *    Spencer: good point, and also makes
    sense to consider what’s changed *    Gorry: measure before and test gently
    during initial deployment. If they’d seen it early on they would have fixed
    it. *    Eric Kinnear: to “actionable” point: fallback strategy and
    end-user pain mitigation is also key to solving it vs. turning it off. *   
    Kireeti Kompella: +1 to Eric’s “fallback plan”, +1 to Colin and Spencer’s
    response, but also generations are sometimes smaller now with faster OS
    updates and microcode, might be smaller than 10 years per generation for
    some cases. Plus, it’s not just “something went wrong” but also taste left
    in your mouth when it went wrong makes you reluctant to retry *    Jake:
    one protocol perhaps mistake to consider is ECT(1)==ECT(0) instead of NECT
    (do what you were doing before, instead of do a new thing equivalent to the
    new thing for an unused codepoint). Would have smoothed things for ongoing
    later attempt to extend.

15:45 15m A Vocabulary of Path Properties (draft-irtf-panrg-path-properties) T.
Enghardt, C. Krähenbühl

    *    Brian: good starting point, make sure service chaining vs. endpoint
    services is clear for service invocation (here we’re talking about service
    chaining). This doc has the problem of context overload for all the terms.
    *    Jake: Will send an email, there are a couple of maybe-examples. *   
    Brian: do we want the questions draft to have a standing forward reference
    to the properties draft or do we want to finish questions doc and point
    only to an abbreviated early version of properties. *    Theresa: path
    properties has an answer to the questions draft. We should be able to
    publish the question’s draft without the properties. *    Brian: Thanks,
    will follow up in a reply to Adrian’s message to the list. *    Spencer:
    what-not-to-do no longer has normative reference to questions’ definition
    of “path” *    Med (from chat): Could just copy/paste the definition of
    path to remove dependency. (Brian and Theresa: +1 each)

16:00 5m Path properties for space-terrestrial networks
(draft-zheng-panrg-path-properties-istn) Shaowen Zheng

    *    discussion to list due to technical audio difficulties

16:05 10m draft-garciapardo-panrg-drkey Juan A. García-Pardo

    *    discussion to list for time considerations

16:15 10m On the subject of trustworthiness S. Dawkins

    *    Brian: we should do an interim before IETF 111. Probably in May.

16:25 5m Gateway trust relationship
(draft-du-panrg-gateway-based-trust-relationship) Zongpeng Du

    *    discussion to list, out of time.

Jen & Brian: will schedule an interim targeting mid-late May/early June,
scheduling poll to follow on list.