CATS WG Meeting IETF 126 (Vienna) 23 July

Chairs: Adrian Farrel, Peng Liu; AD: Jim Guichard

Agenda Bashing

[chairs] New WG milestones added.
[Adrian] Do you know your plan of your travel to IETF127? The answer will help the Chairs to organzie the next meeting.
Voting result: Y:3; N:9; No opinion: 6

WG documents presentations

2.1 - CATS Metric Definition [Kehan Yao]

[Adrian] Two more drafts are related to this metric draft. Do you think we should merge them into this draft or keep them seperately?
[Kehan] Already discussed with some authors from those two documents. We might merge some content from those drafts into the metric draft in terms of operational considerations.

2.2 - CATS OAM Framework [Quan Xiong]

[Luis Contreras] Service is running in virtualized environment, would not be simpler to probe?
[Quan]The CATS OAM just monitors the computing metric reporting, not the metric inside the service instance.

[Boris Khasanov] Why wasn't STAMP considered as an OAM tool? STAMP is more advanced protocol.
[Xiong Quan] The document just provides examples for existing mesurement tools such as STAMP and TWAMP.

[Cheng]Thank you for your work. I have reviewed the document. Sorry that I have not replied to your email yet. However, It seems to me that we might add more OAM mechnisms content to the draft, right now, it looks not so much new information related to CATS in it. Let's make it a useful draft, thank you!
[Quan] ok. Will be happy to do so.

Personal I-D presentations

3.1 Protocols Applicability for Computing-Aware Traffic Steering (CATS) (15 min) [Linda Dunbar]
https://datatracker.ietf.org/doc/draft-yxl-cats-protocols-applicability/

[Mengyao Han] have a draft on BGP-LS, may consider to contribute to this draft. How to consider different protocols to collaborate?
[Peng Liu - chair] This draft is positioned to analyze the protocols not recommend.

[Cheng Li] Need to consult BGP, IGP and PCE experts to join.
[Boris Khasanov] You mentioned YANG model. Why don't consider gNMI or YANG Push?
[Lindar] Thank you, we will check them too.

[Jim - cats AD] I don't want to see the WG jump in the gun. I want to see protocols perspective on requirements for protocols not the pre-assumptions about protocols itself. Also CATS metrics are not finally ready yet so I would like to see them too.

[Jeffrey Haas - IDR chair] Clarify behavior goals and requirements within the WG first, then determine protocol content. Do initial work here, submit to IDR only after maturity — this avoids the premature protocol pitfalls seen with SAV and others.

3.2. Yang Model for CATS (10 min) [Quan Xiong]
https://datatracker.ietf.org/doc/draft-yl-cats-data-model/

[Adrian Farrel] Do you think the document is mature enough now that we could make use of YANG doctor team to review?
[Quan Xiong] Yes, I think so.

[Adrian Farrel] So chairs will ask YANG doctor team for a review.

[Jeffery Haas] A warning that YANG doctor team is for structural and syntax review, not semantic review. Please make sure about that.

3.3.Distribution of Service Metadata in BGP-LS (10 min) [Mengyao Han/Tianyi Huang]
https://datatracker.ietf.org/doc/draft-ls-idr-bgp-ls-service-metadata/

[Sue Hares] It is related to IDR, is this draft necessary for the framework. When you come back to IDR, I'm asking author and CATS WG chairs that the framework needs the protocl extension.
[Adrian Farrel] Protocol comes to other WG for CATS is premature now.

[Jeffrey Haas - IDR chair] Please check the drafts from your list for the common metrics which might be presented in BGP-lS.

3.4.Computing Service Metric Definitions and Operation under CATS (10 min) [Yina Dai]
https://datatracker.ietf.org/doc/draft-zhangb-cats-service-metrics-op/

[Kehan Yao] This document is useful as input to CATS metrics definition draft specifically on operational considerations. But recommend not go separately, since CATS WG should not standardize specific implementations.
[Yina Dai] Sugguest this document go separately.
[Sue Hares] Are you expecting that computing time is synchronized across a network. or you're simply looking for a delta within a networ?
[Yina Dai] The delta time. We are not talking about the time synchronization here.
[Sue Hares] Thanks, your answer helps my work in reviewing IDR drafts.

3.5.Public Service Platform for Computing-Aware Traffic Steering (10 min) [Yina Dai]
https://datatracker.ietf.org/doc/draft-zhangb-cats-cmas/

[Yunze Wei] DAWN BOF on Agent discovery. What's the relationship between this work with DAWN?
[Yina Dai] DAWN is about AI Agent discovery. This work is more on CATS.

3.6. Compute-Aware Traffic Steering for Midhaul Networks (10 min) [Tianji / Luis / Mark]
https://datatracker.ietf.org/doc/draft-lcmw-cats-midhaul/

[Carlos Bernados] This may relate to DMM draft on traffic steering.
[Adrian Farrel] I'm struggling on the archtectural view here. Is this layered arch or a stitched arch?
Are you building ORAN on top of IP?
[Luis] More stitching view.

[Quan Xiong] Any specific input to CATS YANG data model?

3.7. Framework and Applicability of Computation-aware Traffic Steering (CATS) in Optical Transport Networks (OTN) (10 min) [Xiao]
https://datatracker.ietf.org/doc/draft-zhao-cats-otn-applicability/

[No comment...]

3.8. Operational Semantics for CATS Metric Consumption (If time permits) [Mengfei Zhu]
https://datatracker.ietf.org/doc/draft-zhu-cats-metric-semantics/
[Kehan] I discussed this draft with the authors and looks this draft could be useful for metrics semantic definition.

Wrap-up

[Adrian Farrel] Before go to other WGs to show we need protocol X,Y,Z for CATS, come to CATS first to agree on the protocol requirements first.