**Session Date/Time:** 05 Nov 2024 16:30 # tsvwg ## Summary This TSVWG meeting covered updates on drafts and RFCs, working group items, careful resume (including Hackathon presentations), L4S, NQB, SCTP, and future meeting plans. Key discussions revolved around the readiness of Careful Resume for last call, the completion of the NQB consensus call, and Comcast's L4S deployment progress. ## Key Discussion Points * **RFC Publication Updates:** Three new RFCs have been published since the last IETF meeting: Encapsulation Guidelines for ECN, update to the zero checksum SCTPRFC. * **Draft Status:** Updates were provided on the status of several drafts, including Multipath DCP, UDP Options, and NQB. * **Careful Resume:** Discussion on the latest version of the Careful Resume draft, focusing on changes to design principles, transition between phases, BBR considerations, and safe retreat adjustments. * **Careful Resume Hackathon Results:** Presentations on Careful Resume implementations and testing from Fastly and Cloudflare Kiche. * **NQB:** Completion of the consensus call for the NQB draft. * **L4S Deployment:** Comcast provided an update on their low latency deployment, highlighting their progress through crawl, walk, and run phases, and early positive results with AQM enablement. * **Low Latency for Sports:** Comcast exploring the use of L4S to reduce latency in live sports broadcasts. ## Decisions and Action Items * **Careful Resume:** A working group last call will be initiated for the Careful Resume draft on the mailing list after this meeting. * **NQB:** The NQB draft will be updated with the agreed-upon changes and a publication request will be submitted. ## Next Steps * Continue discussions on L4S and SCTP. * Focus on drafts identified as needing review prior to the Friday meeting. * Jason (Comcast) will present the video demonstration on Low Latency improvements immediately after the break. --- **Session Date/Time:** 05 Nov 2024 18:00 # tsvwg ## Summary This TSVWG session focused on L4S operations, interop events, and SCTP with DTLS. Key discussions included the presentation of a video demonstrating the impact of low latency on live sports broadcasts and an update on ongoing L4S interoperability testing. The session also addressed compatibility issues between L4S senders and certain bottleneck types, and the working group discussed how to move forward with DTLS over SCTP, unadopting a previous solution and adopting a multi-document approach leveraging existing IETF standards and potentially new work in the TLS working group. ## Key Discussion Points * **L4S and Live Sports:** A demonstration highlighted the significant differences in latency between various video delivery methods (satellite, OTT, low-latency lab setup), with implications for sports fans and sports gambling. * **L4S Interoperability Testing:** Updates were provided on ongoing L4S interop events, including those held at IETF and Cable Labs. Netflix, Google, and NVIDIA have been actively testing their implementations. A WebRTC interoperability issue was discovered and led to an errata for RFC 88. * **L4S Compatibility Issues:** Discussion continued on the L4S ops draft addressing compatibility issues between L4S senders and certain bottleneck types. * **DTLS over SCTP Solutions:** Multiple solutions were presented but all have different issues. A new approach based on an encryption/decryption component and a separate key management component, the latter leveraging existing IETF standards (DTLS) or work in TLS working group (extended key update). This approach has the potential to avoid IPR concerns. ## Decisions and Action Items * **L4S Ops Draft:** Update the milestone to April 2025 to allow for potential input from IESG review and further deployment experiences. Encourage contributions to the draft. * **DTLS over SCTP:** * **Decision:** Unadopt the existing "DTLS over SCTP" draft (draft-ietf-tsvwg-dtls-sctp-08) as a working group item. * **Action Item:** Editors to produce revision zero of the encryption/decryption and key management drafts by the end of the year for WG consideration. * **Action Item:** Chairs will send a summary of decisions and planned actions in the meeting to the mailing list. ## Next Steps * Continue L4S interoperability testing this week at IETF and in the coming weeks at Cable Labs. * Authors to produce initial drafts of the new DTLS over SCTP proposal for working group adoption. * The working group will review the new DTLS over SCTP proposal once drafts are available. --- **Session Date/Time:** 08 Nov 2024 09:30 # tsvwg ## Summary The TSVWG meeting focused on two main topics: configuring sockets for ECN over UDP and requirements for host to network collaboration signaling for per-packet metadata. The discussion on ECN configuration highlighted inconsistencies across operating systems and the need for better documentation. The discussion on host-to-network signaling sparked a debate about the value of per-packet metadata, security implications, and alternative approaches to congestion control. ## Key Discussion Points * **Configuring Sockets for ECN over UDP:** * Inconsistencies exist in how different operating systems (Linux, Apple, Microsoft, BSD) handle ECN-related socket options. * Documentation for these options is often poor. * The group discussed the potential for adopting the draft as an informational RFC. * There's interest in cross-platform testing and CI to ensure the accuracy of the document over time. * Mentioned the importance of including the contents of the control message which can differ between platforms. * **Requirements for Host to Network Collaboration Signaling:** * The draft proposes using per-packet metadata to improve quality of service for unreliable flows. * Concerns were raised about security implications, particularly exposing metadata that could reveal information about encrypted content. * Discussion on the trade-offs between per-packet vs. per-flow signaling. * The group discussed the potential for stateful vs. stateless schedulers. * Some participants expressed skepticism about the general usefulness of the draft given privacy and security challenges. * Alternative approaches such as focusing on building networks that don't lose packets were also highlighted. * Considerations about the relationship between this work and scone as well as PANRG were also discussed. ## Decisions and Action Items * **ECN Configuration Draft:** * The working group will adopt the draft for informational purposes. * Martin Duke will move the repository to the TSVWG GitHub. * Contributors were asked to file issues and assist with documenting BSD platforms. * The group agreed to confirm the adoption on the mailing list and solicit reviews. * **Host to Network Collaboration Signaling:** * The chairs will talk to the AD about the value and direction of the work. * Continue discussion on the mailing list to address concerns and refine the scope. * Authors of draft to consider framing the discussion about real deployments and benefits. ## Next Steps * Confirm adoption of the ECN configuration draft on the mailing list and solicit reviews. * Continue the discussion on host to network collaboration signaling requirements on the mailing list, focusing on addressing security and privacy concerns and establishing demonstrable benefits. * Chairs will discuss the direction of the signaling work with the AD and report back to the working group.