Minutes IETF106: opsawg
minutes-106-opsawg-01

What: Combined OpsAWG / OpsArea
When: 10:00-12:00 Wednesday Morning session I
Where: Collyer

OpsAWG Section
--------------------

Administrivia  - scribes, minutes, etc.
Tianran / Joe
5 minutes

Layer 3 VPN Network Model
Oscar Gonz¨¢lez de Dios/Qin Wu
Draft:
https://datatracker.ietf.org/doc/draft-ietf-opsawg-l3sm-l3nm/
15 minutes

Auto-assign can be the default behavior
Ignas as operator: Making auto-assign the default will clash with many
deployments; should be an option Oscar: how to make that auto-assignment
happen; leave RT/RD blank or explicitly set auto-assign? Qin: no strong
preference for explicit or implicit auto-assign; document needs to Tim Carey:
We do see in zero-touch the need for the controllers to assign. Leaving the
blank is the auto-assignment. Auto-assignment is done via a profile/policy.
Joe: will send comments to list (on L3SM integration)

A Framework for Automating Service and Network Management with YANG
Qin Wu
Draft:
https://datatracker.ietf.org/doc/draft-wu-model-driven-management-virtualization/
10 minutes

Laurent: Interaction between Service Level and Network level misleading; you
don't decomission the service in the network level; should be within the
lifecycle of the service level Qin: Will shore up service lifecycle

Laurent: Good examples; maybe go for operations and show which models can be
used for those operations Qin: Will work through that Joe: Not on mic; agree
with this comment.  Showing operator use cases and how they map to YANG models
will be very helpful

Yang data model for TACACS+
Bo Wu
Draft:
https://datatracker.ietf.org/doc/draft-ietf-opsawg-tacacs-yang/
5 minutes

Joe: Will follow up with YANG doctors to see if we can do anything with respect
to augment ietf-system Dymytro: Follow up with ietf-system authors to see when
the next version of ietf-system will be published Bo: Will follow up

In-situ Flow Information Telemetry
Haoyu Song
Draft:
https://datatracker.ietf.org/doc/draft-song-opsawg-ifit-framework/
15 minutes

Diego Lopez: Closed loop is misleading; about automatic control
Haoyu: Clarified in draft
Diego: Still not clear, needs a better term to avoid ambiguity
Haoyu: Closed-loop only exists in a given domain
Frank: the scope is large.
China Mobile: useful work and gives some examples.
China Telecom: support the adoption of the work.
Chair: How many of you have read this document? quite a lot.
Chair: How many of you think this is a useful work and the working group could
work on it? still many, 20+.

A YANG Module for uCPE Management
Dmytro Shytyi
Draft:
https://datatracker.ietf.org/doc/draft-shytyi-opsawg-vysm/
10 minutes

Dean: How is this different from RFC8530 [YANG Model for Logical Network
Elements] Dmytro: Not familiar (or don't remember)

Qin: Some similar work in teas working group
Dmytro: Working with the teas work; augmenting it

Dmytro: Corss work going on with ETSI?

Tim (from Jabber): Does the authors intend to use other virtualization
technologies than VMs like Docker, etc Tim (from Jabber): I would also ask that
can't the ETSI NFV SOL drafts be used to solve this problem? If ETSI has solved
the problem why do we need to duplicate this for a very specific problem domain

Sampled Traffic Streaming
Andrew Gray
Draft:
https://datatracker.ietf.org/doc/draft-gray-sampled-streaming/
10 minutes

Frank: Compare to 5476
Ignas: Please schedule a side meeting here in S'pore to discuss this work

In-situ OAM Deployment
Frank Brockners
Draft:
https://datatracker.ietf.org/doc/draft-brockners-opsawg-ioam-deployment/
5 minute

Multiple at mic: useful work, but could use more concrete recommendations. 
That may require more experience before those recommendations can be made It
may be that this document lives and evolves over time as more experience is
learned with IOAM

MUD (D)TLS profiles for IoT devices
Tirumaleswar Reddy
Draft:
https://datatracker.ietf.org/doc/draft-reddy-opsawg-mud-tls/
5 minutes

Eliot to comment on mic, but we had to push this to the list

Ops-Area Section
---------------------
Administrivia - scribes, minutes, etc.
Warren / Ignas
15 minutes

[Short mic time due to long opsawg]

Service Assurance for Intent-based Networking (SAIN)
Benoit Claise
Draft:
https://datatracker.ietf.org/doc/draft-claise-opsawg-service-assurance-architecture/
15 minute

Dean: This is a problem to solve; confusing autonomic and intent-based
Jeff: How do you model data received from CLI, telemetry, SNMP, etc. in a
coherent way Robin: More layers are introduced that add complexity to the
existing layers already present in the IETF

Rajiv: Architecture very useful; from a perspective of config there has not
been an effort to model the assurance aspect and create that hierarchy
 Contain and define service assurance (do you mean network, security, etc.)?
Figure out a hierachy and model that hierarchy
Physical vs. logical awareness is important
Benoit: API is there to augment those sub-service types
Alex Clemm: service assurance and the dependency graph could be split for more
reuse

Open Mic

==================================================

Jabber Log

[00:41:58] chsharp@jabber.org leaves the room
[01:42:40] Meetecho joins the room
[01:50:59] VirtualQueue_FaW78Rka joins the room
[01:55:09] Joe Clarke joins the room
[01:55:11] Timothy Carey joins the room
[01:55:15] Oscar Gonzalez de Dios joins the room
[01:56:04] Lorenzo Miniero joins the room
[01:58:29] <Meetecho> Joe Clarke: yep, this is the Jabber room :)
[01:58:50] <Joe Clarke> You guys are awesome.  Thanks.
[01:59:02] <Meetecho> :D
[02:00:13] <Oscar Gonzalez de Dios> Hi guys!
[02:00:45] <Joe Clarke> Hey, Oscar.  You're up first after the intro.
[02:01:35] adrianolddog joins the room
[02:02:49] <adrianolddog> Anyone in jabber wanting a comment relayed to the
microphone please write "MIC:" [02:03:05] danielking.net joins the room
[02:03:09] <Joe Clarke> Thanks, Adrian. [02:04:22] Simon Pietro Romano joins
the room [02:17:17] <Joe Clarke> Oscar, audio break-in is not working.  you
only have three minutes left, and I want to make sure you get to some of these 
questions. [02:18:16] <Meetecho> Joe Clarke: are you experiencing issues?
[02:18:41] <Meetecho> If you tried talking to him, it may be it got lost
because we lowered capture volume in the room to avoid echo [02:19:12] <Joe
Clarke> Ah.  It's okay now.  He's the only remote. [02:21:40] <Timothy Carey>
MIC We do see in zero-touch the need for the controllers to assign. Leaving the
blank is the auto-assignment. Auto-assignment is done via a profile/policy.
[02:23:33] Luay Jalil joins the room [02:23:41] Andy Bierman joins the room
[02:25:58] <Joe Clarke> Sorry, Tim.  I didn't see this before we cut the line. 
Please follow up on list. [02:26:07] <Timothy Carey> Sure [02:26:21] <Joe
Clarke> I will record your comments in the minutes. [02:30:51] <adrianolddog>
Sorry Tim, I also looked away for a moment [02:40:07] <Oscar Gonzalez de Dios>
@Tim, I'll be glad to follow the discussion on the list [02:41:17] Luay Jalil
leaves the room [02:43:56] Willard Gray joins the room [02:44:21] Willard Gray
leaves the room [02:44:24] Willard Gray joins the room [02:48:24] Oscar
Gonzalez de Dios leaves the room [02:55:13] Lorenzo Miniero leaves the room
[02:59:12] zhenqiang li joins the room [03:00:58] Reshad Rahman joins the room
[03:01:59] <Timothy Carey> MIC I am confused as to how this framework work with
the already adopted Network Telemetry framework draft [03:02:09] <adrianolddog>
got it [03:03:13] <adrianolddog> 2nd in q [03:10:02] Simon Pietro Romano leaves
the room [03:17:38] Sandra Murphy joins the room [03:19:24] <Timothy Carey> MIC
Does the authors intend to use other virtualization technologies than VMs like
Docker, etc [03:20:58] Raymond Ho joins the room [03:21:34] <Timothy Carey> MIC
I would also ask that can't the ETSI NFV SOL drafts be used to solve this
problem? If ETSI has solved the problem why do we need to duplicate this for a
very specific problem domain? [03:22:31] <adrianolddog> queued [03:23:26]
<adrianolddog> sorry @Tim , you got bounced by the chairs because of timing
[03:23:39] <adrianolddog> Please send your comment by email to the list
[03:23:45] <Joe Clarke> Yeah, sorry.  We're running long.  I was hoping we'd
pick up more time on that, but... [03:23:56] <Joe Clarke> Recording comments in
minutes.  Please take the question to the list. [03:24:04] <Timothy Carey> Sure
no worries [03:30:28] chriscroot joins the room [03:38:18] <zhenqiang li> what
is the relationship of this one to the one presented by Song? [03:41:17]
Willard Gray leaves the room [03:42:08] danielking.net leaves the room
[04:02:47] adrianolddog leaves the room [04:03:23] chriscroot leaves the room
[04:03:38] Andy Bierman leaves the room [04:03:49] Raymond Ho leaves the room
[04:03:49] Meetecho leaves the room [04:04:03] Joe Clarke leaves the room
[04:04:03] zhenqiang li leaves the room [04:04:03] Timothy Carey leaves the
room [04:04:03] Reshad Rahman leaves the room [04:04:04] Sandra Murphy leaves
the room [04:04:10] chriscroot joins the room [04:09:44] chriscroot leaves the
room [04:09:50] chriscroot joins the room [04:10:42] chriscroot leaves the room
[05:21:32] adrianolddog joins the room [05:21:42] adrianolddog leaves the room