**Session Date/Time:** 18 Mar 2025 02:30

# ccwg

## Summary
This meeting of the Congestion Control Working Group (CCWG) included updates on various congestion control tools and proposals. A significant portion was dedicated to updates on the BBR congestion control algorithm, including open issues, potential improvements, and the path towards RFC publication.  Presentations were also given on CC-Perf, NEST, and rate limited senders.

## Key Discussion Points

*   **Congestion Control Testing Tools (CC-Perf, NEST):**
    *   Undergraduate students from NITK presented CC-Perf and NEST.
    *   CC-Perf is built on NS3.
    *   NEST is built on Linux namespaces.
    *   Hackathon results were shared, including testing of FQ Codal and FQ PI queue disciplines.
    *   Concerns raised about lack of native QUIC model support in NS3.
*   **BBR Updates:**
    *   Discussion about the bar for publishing the BBR RFC, with a preference for multiple deployments at scale in both QUIC and TCP.
    *   Debate about including ECN support in the initial RFC versus a follow-on document (BBR-BIS).
    *   Open PRs: a correction to drain rate calculations.
    *   Discussion on how BBR should treat ECN signals, particularly ECT(0).
    *   Concerns raised about BBR's suitability for real-time applications due to probe RTT intervals.
    *   Need for clear documentation on areas of BBR that still needs improving.
*   **Rate Limited Senders:**
    *   Update on the draft concerning congestion window increases when the sender is rate-limited.
    *   Discussion on simplifying the rules and clarifying the definition of "rate limited".
    *   Pacing limits should also be considered.
*   **aSearcH:**
    *   Presentation positioning aSearcH relative to other algorithms.
    *   Memory usage analysis.
    *   Test framework described.
    *   Some discussion about algorithmic differences relative to HighStart and variations in RTT measurements.

## Decisions and Action Items

*   **BBR:**
    *   File issues for tests that need to be considered for BBR.
    *   Re-review the BBR document for TCP specific language, so that the document is more transport agnostic.
    *   Consider publishing BBR without ECN support, and address ECN in a follow-on document (BBR-BIS).
    *   Determine as a group which BBR improvements/changes are in scope for "now" versus "next" (BBR vs BBR-BIS).
*   **Rate Limited Senders:**
    *   Re-evaluate the rules proposed in the draft.
    *   Clarify the definition of the term "rate limited" within the document.
    *   Take pacing limits into account.

## Next Steps

*   Authors of various congestion control proposals to address feedback received during the meeting.
*   CCWG chairs to create a list of topics to be considered "now" vs "later".
*   Further discussion and contributions on the CCWG mailing list.