Session Date/Time: 24 Jul 2024 16:30

cats

Summary

The CATS working group meeting covered several topics including use case updates, framework updates, compute modeling and metrics, and applicability to mid-haul networks. Discussions focused on refining existing documents, adding new use cases, defining metrics for computing-aware traffic steering, and addressing security concerns. A new co-chair, Med Boucadair, was announced to replace Adrian Farrel in Dublin.

Key Discussion Points

Use Case Updates:
- Updates to the CATS problem statement, use cases, and requirements document were presented, including new use cases for CDN, cloud rendering, and internet vehicles.
- Terminology changes regarding "service" and "computing service" were discussed, and consensus was reached on updated definitions.
- Deployment testbeds in China for CDN 4K video delivery and cloud rendering were presented, along with preliminary results.
- A use case for high-speed internet for vehicles was introduced, focusing on remote driving scenarios and the need for low latency, high bandwidth, and stable connectivity.
- Discussion on Intelligent Transportation Systems (ITS) use cases, focusing on context-aware navigation and cluster-based MAC protocols, with a focus on the requirements of ADAS.
- Industrial Internet use cases were discussed, focusing on deterministic management of computing power and network resources for applications like industrial robots and virtual cloud termination.
Framework Updates:
- Updates to the CATS framework document were presented, addressing comments from the working group adoption call and identifying areas for further development.
- Discussion on aligning the framework with the use cases being prototyped and experimented with.
Compute Modeling and Metrics:
- Updates to the draft on computing modeling description were presented, focusing on defining computing-related metrics and how to use them in the CATS domain.
- Design principles for metric definition were discussed, including simplicity, scalability, interoperability, stability, and accuracy.
- A hierarchical framework for defining metrics was proposed, with three levels: raw metrics, normalized metrics, and fully normalized metrics.
- Discussion on Kubernetes metrics and architecture.
- A draft on CATS metric definition was introduced, proposing an initial framework for defining metrics.
Mid-Haul Networks Applicability:
- The applicability of CATS to mid-haul networks was analyzed, focusing on the F1 interface in 3GPP terminology and the use of CATS for traffic steering between distributed units (DU) and centralized units (CU).

Decisions and Action Items

Action Item: Kehan to welcome discussion on list.
Action Item: Paul to provide suggested text for merging the ITS use case into the use case document to the authors of the use case document.
Action Item: Photo to review and respond to Carlos Bernal's comment about the industrial requirements draft in the Deadnet working group via email.
Action Item: Ming to add specific text to sections 3.3, 4.1, 4.5, and 5 of the framework document.
Action Item: Working group members to provide feedback on the compute modeling and metrics drafts.
Action Item: Chairs will talk to the AD to discuss the best way to communicate with the O-RAN Alliance.
Action Item: Linda to send a note to the mailing list with a pointer to the IDR draft on the metadata past attribute.
Action Item: Group to consider Security issues.
Decision: The chairs will talk to the AD and work out what the best way to communicate with O-RAN is

Next Steps

Authors to continue refining documents based on feedback.
Working group to discuss and agree on default metrics for CATS.
Explore the interplay between the ORAN orchestration framework the SMO, and how it could interact with the CATS entities
Prepare new versions of drafts for the next IETF meeting.
Push out new revisions more often, in order to provide cleaner reading materials for the drafts.