**Session Date/Time:** 30 Mar 2023 04:00

# grow

## Summary

The grow working group session at IETF 116 focused on updates and discussions related to BMP (BGP Monitoring Protocol). The agenda included presentations on TLV and EB bits, the BMP YANG model and D-path marking, and BGP Local RIB peer address. Key discussion points revolved around versioning, experimental bits, and statelessness in BMP.

## Key Discussion Points

*   **TLV Draft (Paolo Lucente):**
    *   Discussion on bumping the version to version 4 due to non-backward compatibility with existing implementations. The debate centered on the cost of version bumping versus creating a new message type.
    *   Concerns regarding stateless parsing Tl were raised. The idea of adding an "ad path" flag in the header was discussed as an alternative.
    *   Agreement to solicit feedback on version bumping versus a new message type using a mailing list survey.
*   **EB Draft (Paolo Lucente):**
    *   Conflict with IANA allocations for experimental use of TLV types was identified.  The solution is to remove the IANA experiment allocations and move to using eb for experimental statistics.
    *   The group discussed whether the EB draft should apply to the stats message as well.
*   **BMP YANG Model (Camilla):**
    *   Updates to the BMP YANG module since the last meeting were reviewed, including removing "id y" from identity names and adding an identifier for the global network instance.
    *   The need for clear explanations and examples in the draft was emphasized, encouraging wider community input.
*   **D-path Marking (Camilla):**
    *   Discussion on standardizing the basic status codes, though manufacturers' experimental wait states might remain EB bits forever.
    *   Debate on whether to standardize recent codes and the size of the recent code, which is two octets. A call for adoption for this draft was initiated.
    *   Discussion about the best way for a router to encode reasons to reject messages.
*   **BGP Local RIB Peer Address (Fran):**
    *   Proposal to allow the peer address field to be non-zero to identify the source of installed routes.
    *   Debate ensued regarding potential backward compatibility issues and the impact on implementations. Jeff highlighted inconsistencies with the v bit in different record types, and potential workarounds using available bits.
    *   Discussion on the statefulness of the collector vs the router, and a general preference to strive towards statelessness on the router.

## Decisions and Action Items

*   **TLV Draft:** Conduct a survey on the mailing list regarding bumping to version 4 versus creating a new message type.
*   **EB Draft:** Update the draft to reserve old code points or say that a number should be removed.
*   **D-path Marking:** Working group adoption of the D-path marking draft was initiated.
*   **BGP Local RIB Peer Address:** Authors will revise the document to incorporate feedback from the session, including Jeff's comments, and upload a new version.

## Next Steps

*   Authors to follow up on action items and post updated drafts.
*   Continue discussions on the mailing list for remaining open issues.
*   Consider working group adoption for drafts with sufficient progress.