**Session Date/Time:** 29 Jan 2024 15:00

# [IDR](../wg/idr.html)

## Summary

The IDR Working Group session focused on the status of several drafts, with a particular emphasis on preparing the Color-Aware Routing (CAR) and Color-aware Transport (CT) drafts, as well as the BGP Send Hold Timer draft, for Working Group Last Call (WGLC). Key discussions revolved around the technical details of CAR's Local Color Mapping Extended Community (LCM EC), the SRv6 data plane aspects of CT, and the proposed BGP Send Hold Timer mechanism. Several action items were assigned to resolve lingering technical comments and prepare the drafts for final review. The multi-next-hop discussion was deferred to a future interim.

## Key Discussion Points

*   **Working Group Last Call (WGLC) Schedule:**
    *   WGLC for `draft-ietf-idr-bgp-car` and the `draft-ietf-idr-bgp-ct` series (base CT and CT SRv6) is planned from February 1st to February 22nd. This timeline is contingent on timely edits and resolution of open issues.
    *   WGLC for `draft-ietf-idr-bgp-send-hold-timer` will follow after IANA code point allocation.
*   **Aid from Area Directors and Other WGs:**
    *   The chairs will request assistance from the new Area Director, as well as Spring and BEST Working Group Chairs, for final reviews of the CAR and CT drafts, particularly regarding the need for an interoperability draft.
*   **Status of Other Drafts:**
    *   `draft-ietf-idr-mbbg-extension-for-map6` is in active discussion.
    *   `draft-ietf-idr-multi-top-attribute-10` is adopted but anticipated to require extensive discussion.
    *   `draft-slyz-idr-bgp-generic-metric-aip` has interest, awaiting author engagement.
    *   `draft-ietf-idr-multi-next-hop-attribute` discussion deferred to the next interim.
    *   `draft-ietf-idr-segment-routing-te-policy` needs comments resolved.
    *   `draft-ietf-idr-ent-label` awaits a second implementation.
*   **Color-Aware Routing (CAR) Discussion (`draft-ietf-idr-bgp-car`):**
    *   **Shepherd and Directorate Reviews:** Approximately 10-12 technical and 40 editorial issues from the first WGLC and directorate reviews (Routing, Ops, Security, TSV) have been addressed in versions 3-5.
    *   **Interoperability:** Cisco has an implementation, and Arcus's is in progress.
    *   **Local Color Mapping Extended Community (LCM EC):**
        *   Detailed explanation of how LCM EC carries color/intent across color domain boundaries, with values being rewritten by local policies at peering points.
        *   **Type 1 Routes:** The NLRI carries the color in the originating domain. The LCM EC is attached and rewritten at domain boundaries.
        *   **Type 2 Routes:** The IP prefix is the route key, with the color carried in the LCM EC from origination, also rewritten at domain boundaries.
        *   **LCM EC vs. Color EC:** Discussion clarified that LCM EC represents the route's end-to-end intent, while the Color EC (RFC 9256) primarily guides next-hop resolution, with different precedence rules.
*   **Color-aware Transport (CT) Discussion (`draft-ietf-idr-bgp-ct` and `draft-ietf-idr-bgp-ct-srv6-data-plane`):**
    *   **Shepherd and Directorate Reviews:** Version 22 of the base CT draft has addressed comments from KO, G, Jonathan, Med, Bo, and Magnus. Mostly editorial with some re-organization.
    *   **Normative References:** Removed normative reference to the Multi-Top (M&H) draft and clarified informative references.
    *   **CT SRv6 Data Plane (`draft-ietf-idr-bgp-ct-srv6-data-plane`):**
        *   This draft was separated from the base CT draft for clarity. It employs the SID stacking approach using SRv6 endpoint behaviors (RFC 8986) and `draft-sallaty-srv6-inter-domain`.
        *   **SID Stacking Example:** A detailed inter-AS scenario with P1 and P2 demonstrated B.ENCAP.S and B.REPLACE.B6 SIDs in packet flows, including the roles of ASBRs.
        *   **Data Plane Questions:** Concerns were raised regarding source addresses in packet diagrams, the carrying of potentially unnecessary SIDs end-to-end, and the use of implicit null (label 3) in the NLRI for SRv6 CT, particularly in MPLS interworking/migration scenarios.
        *   **SRv6 Transposition:** The draft sets SRv6 transposition length and offset to zero to avoid polluting the MPLS label field (RFC 8277).
        *   **Color Prefix Routing (CPR) Support:** CT constructs (resolution scheme) are applied to the IPv6 family with Color Extended Communities for SRv6 locators, without direct CT family involvement.
        *   **Dependency on `draft-sallaty-srv6-inter-domain`:** A question was raised about proceeding with WGLC for CT SRv6 when its underlying Spring dependency draft (`draft-sallaty-srv6-inter-domain`) is not yet adopted. The chair indicated a "shepherd's choice" to proceed for forward progress, acknowledging a potential need for a second WGLC.
*   **BGP Send Hold Timer Discussion (`draft-ietf-idr-bgp-send-hold-timer`):**
    *   **Problem:** Remote BGP peers with a zero TCP receive window can prevent successful delivery of BGP messages (e.g., KeepAlives), leading to prolonged open but inactive connections.
    *   **Proposed Solution:** A mandatory `send-hold-timer` that expires if no BGP message is successfully delivered for a configured duration (default 8 minutes).
    *   **FSM Changes:** The timer starts upon entering the `Established` state and is reset upon successful delivery of any BGP message.
    *   **On Expiry:** A `Notification` message with a new BGP error code "Send Hold Timer Expired" is sent (new in the latest version), and the local system logs an error, releases resources, and drops the connection, similar to a regular Hold Timer expiry.
    *   **Implementation Status:** Four existing implementations.

## Decisions and Action Items

*   **Decisions:**
    *   Proceed with WGLC for `draft-ietf-idr-bgp-car`, the `draft-ietf-idr-bgp-ct` series (base CT and CT SRv6), and `draft-ietf-idr-bgp-send-hold-timer`.
    *   Defer discussion of `draft-ietf-idr-multi-next-hop-attribute` to the next interim meeting.
*   **Action Items:**
    *   **Chair (Sue):**
        *   Request early IANA code point allocation for the BGP Send Hold Timer notification.
        *   Host chair Zoom discussions this week to resolve lingering technical issues for CAR and CT drafts.
        *   Open new GitHub issues for checking `draft-ietf-idr-bgp-ct` v22 against latest comments.
        *   Follow up with Koton, Swadesh, K. Raj, and Nats on GitHub issues.
    *   **Nats:**
        *   Send corrected slide for CT SRv6 (regarding source addresses in packet diagrams).
    *   **Swadesh:**
        *   Resend the example detailing MPLS/SRv6 interworking challenges with implicit null labels to the mailing list.
    *   **K. Raj:**
        *   Work out an example to clarify how ASBR2 distinguishes P2's MPLS support when an implicit null label is used, and share it on the mailing list/GitHub.
        *   (Joint with Nats): Provide reference to the relevant section in the CT/CT SRv6 drafts for the implicit null label discussion.
    *   **CAR/CT Authors/Editors:**
        *   Publish updated versions of CAR and CT drafts, incorporating latest comments and clarifications, by Wednesday evening (anticipating WGLC start Thursday).

## Next Steps

*   Resolve all outstanding GitHub issues for CAR and CT drafts this week.
*   The chair will request an early IANA code point for the BGP Send Hold Timer notification.
*   Initiate Working Group Last Calls for `draft-ietf-idr-bgp-car`, `draft-ietf-idr-bgp-ct`, `draft-ietf-idr-bgp-ct-srv6-data-plane`, and `draft-ietf-idr-bgp-send-hold-timer` as soon as drafts are finalized and the IANA code point is assigned.
*   Participants are encouraged to review `draft-ietf-idr-bgp-ct-srv6-data-plane` and provide feedback, particularly on any functional problems, acknowledging its dependency on `draft-sallaty-srv6-inter-domain`.