Markdown Version | Session Recording

Session Date/Time: 09 Nov 2023 08:30

tvr

Summary

The TVR working group met to discuss the progress on requirements, use cases, and data models for Time-Variant Routing. Key topics included the requirements document update, use case document review, a YANG model for schedule attributes, and proposals for ALTO-based solutions and IETF protocol extensions for TBR. The group debated the merits of standalone vs. augmented YANG models and whether to include TBR-specific information in routing protocols.

Key Discussion Points

• Requirements Document:

  • Updates included merged content from other documents and expanded terminology.
  • Discussion on how to classify requirements (core vs. use-case specific).
  • Emphasis on short-term, terrestrial use cases alongside long-term LEO/deep space applications.
  • Security considerations and need for verification/identity management of schedules were raised.
  • The use of a controller was discussed, with mention of PC.
  • Concerns were raised about the scalability of the control plane with a large number of nodes.

• Use Cases Document:

  • The "mobile devices" category was renamed to "dynamic resusability".
  • Network topology changes driven by traffic patterns and predictable mobility (e.g., trains, ships) were added as examples.
  • A problem statement section was added, outlining schedule definition, distribution, and execution.

• YANG Model:

  • The design includes a common schedule module, a TBR node module (device model), and a TBR topology module (service model).
  • The discussion centered on the approach of using either a stand-alone module or augmenting existing IETF topology models.
  • Concerns were raised about augmenting the existing models and implementation/deployment dependencies.
  • Questions about whether to model general link availability schedules and whether power state is the right attribute in the YANG model.
  • The importance of synchronization was highlighted.
  • There was significant discussion about whether it should be a stand-alone YANG model versus augmentation of existing models. Some felt it should be augmentations of models such as T-Topology. Others wanted the module to be immediately deployable and concise. A third option, for a common grouping, was also suggested.
  • There was discussion about scope of the module and whether its purpose was for simply moving data or for the scheduling of things in general.
  • It was suggested to make use of existing groupings and definitions, so the models defined here can be easily used by any other models for augmentations.

• ALTO Extension:

  • The proposal leverages ALTO for off-path advertisement of schedule changes to applications.
  • An assessment of ALTO's cost calendar functionality against TBR requirements was presented.

• IETF Protocol Extensions:

  • The proposal involves extending IETF routing protocols (IS-IS, OSPF) to distribute schedule information.
  • Concerns were raised about using routing protocols as transport mechanisms for management information.
  • Backward compatibility with partial deployment using FlexAlgo was discussed.
  • Concerns about the need for synchronization for routing domains with IETF Protocol extensions were raised.

Decisions and Action Items

• Requirements Document: Authors will clean up conformance language. The authors will address the need for operational and security considerations. Authors will consider feedback about how requirements are summarized.
• Use Cases Document: Consider working group last call if no major concerns arise on the mailing list.
• YANG Model: Working group to continue discussing on the mailing list the best way to approach the Yang model: stand-alone volume or augmentation. Authors to review and align the terminology to include existing groupings definitions to make it easier to be used by other models. An analysis of augmenting existing models should continue. If standalone is chosen, all the information in the grouping must be included.
• IETF Protocol Extensions: Authors to capture suggestions from the meeting on the mailing list.

Next Steps

• Continued discussion on the mailing list regarding requirements, use cases, YANG model approach, and IETF Protocol Extensions.
• Potential interim session to focus on control plane implications of TBR.
• Authors to make changes to the draft documents based on the feedback received.
• Determine if the use case document is ready for a working group last call.