Minutes IETF117: opsawg: Wed 20:00
minutes-117-opsawg-202307262000-00
Meeting Minutes | Operations and Management Area Working Group (opsawg) WG | |
---|---|---|
Date and time | 2023-07-26 20:00 | |
Title | Minutes IETF117: opsawg: Wed 20:00 | |
State | Active | |
Other versions | markdown | |
Last updated | 2023-08-03 |
What: Joint OpsAWG / OpsArea
When: 13:00 - 15:00 Wednesday Session II, July 26, 2023
Where: Plaza B
OpsAWG Section
Administrivia - scribes, minutes, etc.
Tianran / Joe / Henk
5 min
Two IPFIX extensions:
https://datatracker.ietf.org/doc/draft-ietf-opsawg-ipfix-srv6-srh
https://datatracker.ietf.org/doc/draft-ietf-opsawg-ipfix-on-path-telemetry
Thomas Graf
10 min
Joe: How many have read thes drafts?
- Few hands.
- No questions/discussion.
Digital Map drafts:
Modeling the Digital Map based on RFC 8345: Sharing Experience and
Perspectives
https://datatracker.ietf.org/doc/draft-havel-opsawg-digital-map/
Olga Havel
10 min
D: When referring to technology-specific layer do you mean optical?
Olga: Yes, amongst others
Danielle: Lots of work being done in CCAMP.
Olga: We will look at all this work
Italo: I agree with Daniele. Also look at the work done in TEAS
(RFC8795). In CCAMP we have recently worked on a solution to report a
multi-technology network in a single topology instance
Joe Clarke (as contributor): Why call it "Digital Map"? "Digital Twin"
seems to capture the idea of a holistic approach to modelling services
and topology and exposing them.
Olga: How to combine these different aspects of digital twin. All goes
back to the core, basic layered topologies. It will link all the
functional models and data of the digital twin together.
Joe: Isn't that just applicable as Digitial Twin.
Benoit: there’s topology, the digital map, and the digital twin (a
replica of the network to play with offline simulation). On top of
topology, you might need traffic, flows, and more ... to realize the
digital twin. The digital map is just a step towards digital twin.
Carsten: How to test a topology? That problem has been studied and there
are emulators that have been doing the job of simulation.
Med (from chat): Joe is correct that this is not specific to digital
twin
Olga: It can be used outside of the Digital Twin as well
Vishnu Beeram (from chat): Is draft-havel-opsawg-digital-map intended to
be an informational document (is the current intended status an
oversight)?
Benoit (from chat): Vishnu, it depends whether we want to specify some
of the solutions in that document. With the actual content,
informational seems more appropriate.
Some Refinements to RFC8345 (Network Topologies)
https://datatracker.ietf.org/doc/draft-davis-opsawg-some-refinements-to-rfc8345/
Nigel Davis
5 min
Jan Linblad (from chat): Note that removing/deviating out the source and
destination containers, then augmenting them back in will not give you
compatible JSON (nor XML) representations, contrary to what was stated
in the presentation
Alex Clemm: Interesting presentation, not read it yet. Goal is to
simplify things. Having multiple alternatives in the model adds
complexity.
Nigel: Don't want to take away existing capability. Want to find a way
to evolve the model from where we are to where we need to get to.
Alex: I'll review the draft.
Ahmed: Haven't read the draft yet. Your slides showed the YANG tree. Is
link-id the only mandatory field.
Nigel: yes, we might even replace it with tp-id references
Ahmed: My worry is that if everything is optional, then what does it
mean when lots of fields are configured.
Nigel: Relying on current RFC8345 YANG version.
Ahmed: Network Markup Language => Open event forum ... won't have
differnet type of links but the notion of "group". Alternative solution
is to group unidirectional links for point to multipoint/bidirectional
links
Nigel: This solution allows you have link groups without specifying.
Italo: For bidirectional links, 2 links uni-directional, between the
same two TPs, works perfectly. Do not know why this approach is
necessary. A similar approach could work for p2mp. For multipoint, you
can simplify the reachability information but reporting stats (delay,
availability) might end up being on the uni-directional links: can't
avoid the N² problem.
Nigel: At an extreme, yes you end up with that. Rarely that complex,
often they have similar properties. Groups of points may give
flexibility, but with less complexity. Can describe all complexity using
groups.
Italo: It seems you want to compact the
.
Joe: [Chairs hat on] Sounds like you have some action items (Italo,
Ahmed, Alex) and you want more iterations on this draft. Encourage more
discussion.
A YANG Data Model for Intermediate System to intermediate System (IS-IS)
Topology
https://datatracker.ietf.org/doc/draft-ogondio-opsawg-isis-topology/
Oscar González de Dios
10 min
Olga: The example of limitation in the ietf-network-topology module for
digital map is link between the areas. For that reason we cannot model
OSPF areas as networks. Would Oscar model OSPF areas as networks?
Oscar:We need to look at the performance when we decide how to model
networks. One network, one domain. Brazil: 25000 router, into a single
domain ☹. Otherwise we can discuss
Joe (as co-chair): How much is this work depending on the other drafts?
Oscar: we will together. It could be atomically consumed but make more
sense to be part of the full picture
Rob (from chat & then mic): Not trying to send this somewhere else, but
there may be enough overall work here to have a separate WG. But it
would be good to present in RTGWG and LSR just to ensure that those
folks are aware of what you are doing, and they may want to help
A Policy-based Network Access Control
https://datatracker.ietf.org/doc/draft-ma-opsawg-ucl-acl/
Qiufang Ma
10 min
Poll to see whether WG is interested to adopt the draft: "Is there
interest to adopt this work"
Joe: Raised hands represent "yes"
(discussion before end of hands poll)
Gargi Adhav: Is this implemented somewhere?
Quifang: yes we have implemented this.
Gargi: Will this policy be on the application server or on the devices?
Quifang: On the network controller.
Gargi: Would require inspecting
Poll Result
Hands raised: 27 "raise hands if you would like to see adoption"
Hands not raised: 04
???: Name of policy?
Quifang: Time-based policy,
An Update and Next Steps for the Attachment Circuits Specifications
https://datatracker.ietf.org/doc/html/draft-boro-opsawg-teas-common-ac
https://datatracker.ietf.org/doc/draft-boro-opsawg-teas-attachment-circuit/
https://datatracker.ietf.org/doc/html/draft-boro-opsawg-ntw-attachment-circuit
Mohamed Boucadair
15 min
Joe: Will check on whether a liaison reply back to ORAN is requried.
Med: I think that it might just have been "for information", but sending
a reply may still be nice.
Joe: We will check. Action on Joe/chairs to check and formulate a reply
if needed.
No comments/questions.
Poll to see whether WG is interested to adopt the draft: "raise hands if
you would like to see adoption"
Poll Result
Hands raised: 18
Hands not raised: 1
75 people in the room/online.
Incident management for network service
https://datatracker.ietf.org/doc/draft-feng-opsawg-incident-management/
Qin Wu
10 min
Alex Clemm: Why do you think this fits into here in OPSAWG in the IETF.
Why not is OSS layer or TMF.
Qin: We are defining network level incident management, and this is
designed to help with OSS.
Daniele: You are speaking between the controller and BSS. ...
Qin: We have some multi-level hierarchy controller. This incident
management could work with multi-layer architectures.
Joe: Please take remaining questions to the list.
A Data Manifest for Contextualized Telemetry Data
https://datatracker.ietf.org/doc/draft-ietf-opsawg-collected-data-manifest/
Jean Quilbeuf
10 min
Victor Lopez (Nokia): would be interested in contributing to the work.
Joe: Leave this to be discussed on the list. Have you considered
contacting Netmod.
Applying COSE Signatures for YANG Data Provenance
https://datatracker.ietf.org/doc/draft-lopez-opsawg-yang-provenance/
Diego R. Lopez
10 min
Juan Cardona (from chat, during slide 3): Static mount, please
Rob: very interesting problem. How do you do that when you will be
pushing bigger chunks data? Have you looked at YANG metadata?
Diego: could be part of annotation, might be loophole wrt recursion.
Jean: right place could be the YANG-push notification message, to have
signed telemetry.
Balazs: Did you consider usign YANG instance data file format, for
sending or offline storage.
Diego: Will check the different choices. Will discuss on the list.
Per Andersson (from chat): I missed the queue on the yang data
provenance but would be interesting to have multiple signatures for the
data with e.g. COSE_Sign instead of COSE_Sign1, for e.g. endorsements
as Diego mentioned or for approvals of changes before application to a
datastore as Diego mentioned.
It is interesting work.
A Minor Update to Finding and Using Geofeed Data
https://datatracker.ietf.org/doc/draft-ymbk-opsawg-9092-update/
Randy Bush
5 min
Warren: Geolocation works for IETF meetings because we publish it.
Joe(as a contributor): I think adopting is the next step.
Rob (from the Chat): Regarding Randy's work, I missed part of the
presentation, but flicking through the slides I think that this looks
like a relatively minor enhancement and I'm keen that we update
docs/guidance to improve them based on operational experience, after
all, we are in the OPS area :-)
Joe (as Chair): next presentation will be the last. Apologies to Luis
and Marisol, they will be prioritized for the next meeting.
Green Networking Metrics
https://datatracker.ietf.org/doc/draft-cx-opsawg-green-metrics/
https://datatracker.ietf.org/doc/draft-clemm-opsawg-pam-ipfix/
Alexander Clemm
10 min
Daniele: Have you considered that part of this work might fall on the
inventory work starting tomorrow.
Alexander: We define the metrics, not the context of the metrics. This
work might be reused by the inventory effort.
Rob: This in an Ok home to do this kind of work. Inventory not related
to these aspects yet.
Joe (as chair): You mentioned that discussion occurred on another
mailer. Is OPSAWG the right place?
Rob: Right place for now. People interested in green networking might be
new to IETF and need guidance.
Joe (as chair): get people to comment on OPSAWG mailer in order to
adopt.
Laurent: How do we manage the fact that green aspects are cross-domain?
Rob: ...
Alexander presents draft-clemm-opsawg-pam-ipfix
(Cut because of time)
=== Exceeds time ===
A YANG Data Model for Network Diagnosis by scheduling sequences of OAM
tests
https://datatracker.ietf.org/doc/draft-contreras-opsawg-scheduling-oam-tests/
Luis M. Contreras / Victor Lopez
5 min
Lifecycle Management and Operations:
https://datatracker.ietf.org/doc/draft-palmero-opsawg-ps-almo/
https://datatracker.ietf.org/doc/draft-palmero-opsawg-dmlmo/
Marisol Palmero
10 min
Ops-Area Section
Administrivia - scribes, minutes, etc.
Warren / Rob
10 minutes
Rob is stepping down from OPS A in March
Benoit is setting up a side meeting on Digital Map (see
https://www.ietf.org/mailman/listinfo/digitalmap-yang)
Warren calls EOM
Open Mic