Session Date/Time: 21 Jun 2023 14:00

CORE

Summary

This interim meeting of the CORE Working Group focused on the progress and next steps for the core-sid and core-comi drafts, both of which require further working group last calls due to recent revisions and simplifications. Discussions also touched upon the non-traditional-responses draft, where an architectural split into informative and normative documents was proposed, along with a new option concept.

Key Discussion Points

CoAP YANG SIDs (`core-sid`)

Status: The draft was pushed back from ISG processing to the working group, necessitating a new Working Group Last Call (WGLC).
Current Issues:
- Requirement for status information in the YANG SID allocation files.
- The pyang tool used for generating examples was found not to output the complete strings (e.g., /input and /output for RPCs) and missed some tree points for SIDs.
Proposed Fixes:
- Run a patched pyang using the existing SID file as input to avoid renumbering previously allocated SIDs.
- Ensure all relevant SIDs are included, not just for data items, but also for specific points in the data tree (e.g., /rpc/system/1717/*).
- Address editorial aspects to clarify that alternative SID allocation processes are acceptable, provided they meet the objectives outlined in Section 2 of the document.
Timeline: Process exception edits and pyang work to be done in parallel. Prepare draft -21 for discussion on the mailing list and potentially the next interim. Aim for a WGLC to be completed before IETF 117, allowing for further discussion at IETF 117 before resubmission to ISG.

CoAP Management Interface (`core-comi`)

Status: Following feedback from the IETF London hackathon regarding complexity, a significant simplification has been implemented.
Simplification Approach: A non-radical solution was adopted, retaining GET, PUT, and DELETE operations for whole data access, but not introducing FETCH and PATCH for individual tree parts in this draft.
Current Draft: The branch incorporating these changes has been merged and submitted as draft -13, available for discussion on the mailing list.
Wildcard Checks:
- Verify if the pyang changes required for core-sid also impact core-comi's SID file examples.
- Check if the replacement of rc_yang_data with sx_structure (as done in core-sid) is also needed in core-comi.
Timeline: Aim for a WGLC approximately concurrent with core-sid, or even earlier if outstanding questions are quickly resolved. Discuss outcomes at IETF 117 before submitting to ISG.

Further Outlook (Yang Library and Binary Data Formats)

YANG Library: The document is nearing completion, but a check on its complexity level is warranted.
Binary Data Formats for YANG: There is growing interest in defining a way to switch existing YANG models from text-based data representations (e.g., text-based IP addresses, dates) to more efficient binary formats within CBOR. This is anticipated to be a significant design effort, requiring careful consideration of how new binary formats are introduced, and how evolution and differing levels of adoption are managed across deployments.

Non-Traditional Responses

Context: The discussion was noted to be missing a key contributor, Christian, but some preliminary points were covered.
Architectural Approach: A sense of those present indicates support for a unifying architectural approach. It was suggested that the document could be split:
- An informative document describing the broader architecture of CoAP beyond the HTTP blueprint, particularly concerning innovations like multicast responses and observe notifications.
- One or more normative documents defining specific options (e.g., num-additional-responses, multicast-timeout), which currently serve more as illustrations in the draft. These normative options would require more thorough review, implementation experience, and potentially different timelines.
Terminology: Discussion around potentially confusing terms such as "multicast responses" (suggesting "multicast sent response" as clearer) and the broader category of "non-traditionally transmitted requests" (e.g., "configured requests" vs. "phantom requests").
New Option Idea (from Christian): A concept for a new option was brought up, similar to multicast-timeout but providing a continuous "send multiple responses until deactivated" mechanism. This would enable a privacy-preserving scenario involving chained proxies with OSCORE, where a generic instruction to forward multiple responses is preferred over explicit timeout values. This would need a corresponding deactivation mechanism, possibly drawing parallels with how the observe option deactivates requests.

Decisions and Action Items

Decision: Proceed with a Working Group Last Call for the core-sid draft (expected to be a 2-week WGLC).
Decision: Proceed with a Working Group Last Call for the core-comi draft, aiming to run it in parallel with core-sid.
Action Item: Shepherd Michael Richardson (for core-comi) to be informed of today's discussion points.
Action Item: Initiate discussion on the non-traditional-responses draft on the mailing list, incorporating existing review comments (e.g., from Marco).
Action Item: Further explore the splitting of the non-traditional-responses draft into informative architectural guidance and separate normative option definitions.

Next Steps

Continue discussions on core-sid, core-comi, and non-traditional-responses on the CORE WG mailing list.
Schedule the next interim meeting, tentatively aiming for two weeks.
Plan for core-sid and core-comi WGLCs to conclude before IETF 117.
Consider a potential side meeting at IETF 117 for further in-depth discussions on these topics, especially non-traditional-responses and new options, ensuring remote participation is supported.
Monitor progress on attacks-on-coap and DNS related topics for inclusion in future interim agendas.