Minutes IETF98: opsawg
minutes-98-opsawg-00

Welcome to Ops Area

Agenda slide

Benoit request from Ralph to present CASM is the second item

Benoit

- transition in the Ops Area
Joel steppng down.

- it was a fun ride

somehow a good duo

trying to do not to not that much damage

years of  service I didn’t understand before joining the IESG

Joel

Warren

Benoit - CASM bof

Ralph Droms presenting - CASM bof

    non-WG forming

    Abstract

needs work on scope and definition

Adrian Farrel - followup question can the scope be defined?

Next presentation
David Black - YANG in Redfish DMTF

Converged Infrastructure

Joe White speaking as part of presentation

    What is Redfish?


Ideas in parrallel with RESTCONF

In principle you could take the CSDL files and expose them via an
outbound interface.

john leung speaking

kent  - this work predates that standardization of netconf

Benoit - very nice, what is your ask?

    Expertise on the YANG models?

David Black all of the above interest help and advice.

Joseph White - minimum for basiline switch as our first profile.

Joe Clark - OpenConfig YANG catalog is a place to collect the catalog.

Benoit - thank you, let me know if you need more contacts. Lets invest
time in doing the right thing.

Next presentation SIDRops - Chris Morrow

    First actual meeting

    4 drafts being worked on

    and the charter is approved.

Joel Jaeggli -

    Protocol works goes to IDR

    We get to mature and operationalize the results.

Open mic

Adrian Farrel - NETMOD vs OpsArea documents we just need to re-align the
works

Benoit - do you believe the clasification document in netmod is good to go?

Adrian - yes

----

OPSAWG agenda -

Introduction - Ignas Bagdonas

Tianran Zhou

WG status slide

Agenda slide

first presentation Zhenqiag Lu -export BGP community information in IPFIX


Ignas Bogdonas - comments on messages size and type

The size of the BGP message will be much larger than this soon so we
should consider that

On the community types, extended communities are used all the time

Regarding wide communities you probably don’t need to be concerned about
the econding

Keep it in one document

Wide communities - are probably a bit special.

Rudiger Volk Deutsche Telecom

   BGP messages will get larger, not sure what concern needs to be
taken about the size.

    Need to look at where the implmentations need the space for storing
the stuff.

    on linecards datastructures on linecards will blow up

    protecting ipfix message size and the load on the transmission side
may bbe a consideration.

Joel Jaeggli -

    There are cases where you don’t have recourse to this information,
that’s fine.

Ignas - who as read this document?
    a few hands
    generally postive

Next presentation - Elliot Lear  - MUD

    normatively uses netmod acl model - get on with it

Comments slide

More issue
    Description is not internationalized - change to url


Will call the question on the iesg list

Should mud url be used for more than just mud
    e.g. anima

Tim Polk - did want to understand if that problem could exist - yeah
there is this sittuation (multiple services to exist)

Kent - would it be possible to dole these out

Elliot - where device is hacked we don’t want to to be putting out new
certs. we want the assertion to come from the manufacturer.

Mahesh - Cisco -  the model (acl) as defined doesn’t work for cisco

    that’s why the model (acl) is stuck

Elliot - you’ll commit to doing the work?

mahesh - yes

    based on the discussion we’ve had

Eliot - would would very much appreciate reviews.

next presentation - the lifecycle of mud thorsten dahm

Kent - security considerations text - worse thing that can happen is
substantially worse then described

Eliot - thanks for doing the drafts

    Potential for a bcp leave these documents open for a while until the
base model is done

Ignas - show of hands, who as read these documents?

    few

    will take to list


Next presentation network artificial intelligence - druv dhudy

Adrian Farrel - like your 30k view archtectural picture

    the two other arrows on the slide enhanced pce slide are challenge
(how much data is that)

Joe Clark - maybe I’m really jaded on artificial intelligenice. it
sounds like that part of your plan but what you have have here is what
do we think generall about analytics

? - we want the use cases bcuase the protocol may be different based on
each of them

? - more tangible proposal on how to proceed

Pushith Bragary - are you thinking of working with a specific control -
maybe not useful for opendaylight but others may be more useful, storage
should be considered also.

Ignas two comments

    seems like a set of proscriptions or descriptions

    seems like more detailed design rather than archtecture


    is the ietf the right place for the operational feedback?

Next presentention - carrier wifi calling deployment considerations -
byju pularikkal

? - it’s useful compared to reading the 3gpp specifications

Eliot Lear -  you have to actually have the normaitve references for
hotspot 2.0

Ignas who has read the document

    a few hands

Eliot and here’s why? onboarding process

Sri cisco - if there’s not much interest in the working group maybe ad
will sponsor it.

Open mic