**Session Date/Time:** 21 Jul 2025 12:30 # rtgwg Meeting Minutes - IETF 123 ## Summary The RTGWG meeting at IETF 123 covered a wide range of topics, including updates on existing drafts, discussions on new proposals, and a poll on BFD/VRRP solutions. Key discussion points included the Destination Source Routing draft, a YANG model for IPV6 Neighbor Discovery, egress protection for SRv6 paths, integration of BFD with VRRP, lightweight routing using LLDP, an SRv6 next-hop YANG model, enhanced ECMP for AI clusters, microloop prevention, lossless data transmission in WANs, and adaptive routing in SCAP networks. Several drafts were suggested for further review in other working groups. ## Key Discussion Points * **Destination Source Routing:** The draft is nearing completion and the authors are requesting a last call. Implementations exist in FutureNet, OpenWRT, and Cisco. Hackathon testing showed RFC 8028 compliance. Concerns raised about comparisons to SRv6. * **YANG Data Model for IPV6 Neighbor Discovery:** The name has been changed back to IETF ND to align with the draft title. There was discussion about the scope of the model and whether it should cover all ND functions not defined in other IETF models. It was decided that the draft should be reviewed by the SIXMAN working group, and a proposal was made that the document be moved to SIXMAN. * **SRv6 Egress Protection:** This draft focuses on providing faster protection for egress nodes and links in SRv6 paths. It includes protocol extensions to IS-IS and OSPF. Questions were raised about why this work isn't in SPRING or LSR and whether it applies to BGP-based services. Jim Guichard required copying SPRING and LSR on the working group last call. * **BFD and VRRP:** Two competing drafts address integrating BFD for faster failover in VRRP. One draft proposes a "critical backup" to reduce convergence time and network overhead. The other draft proposes using BFD for multipoint networks (RFC 8562) and using BFD demand mode. There was disagreement on the need for changes to VRP and the implementation of BFD. A poll was taken to determine if the group should move forward with one solution. * **Lightweight Routing using LLDP:** The proposal introduces extensions to LLDP to advertise IPv6 and IPv4 prefixes. Concerns were raised regarding STP bypass, security considerations, L2 protocols installing L3 routes, protocol complexity, and the limited scope of multicast addresses in Ethernet. The security considerations were with regards to MACSEC. The overall sentiment was largely negative. * **SRv6 Next Hop YANG Model:** Discussion of an SRv6 next-hop YANG model. It was noted that some of the work proposed could potentially be included in existing SRv6 related YANG models currently in SPRING. * **Enhanced ECMP for AI Clusters:** Presentation of two enhanced load balancing solutions for AI data center networks, ingress group and egress group load balancing. Question raised that the assumptions of the traffic model did not hold for the MOE model, where traffic is unevenly distributed. * **Loop-Free Routing by Microloop Implementation:** The draft describes implementation methods of segment routing for microloop avoidance in different scenarios. Question raised about the overlap with RFC 5715. There was another concern of whether implementation details should be documented. * **Use Cases, Requirements, Framework for Implementing Lossless Technicals in Wide Area Network:** The presentation outlined the requirements, a draft framework, and use cases for lossless transmission over WAN and DCN interconnections, noting limitations of existing protocols like TCP and PFC. A question was raised about whether a Layer 4 solution or FEC could be used instead. * **Fully Adapted Routing Internet to Scale Up Networks:** Use of BGP for adaptive routing in SCAP networks was discussed. ## Decisions and Action Items * **YANG Data Model for IPV6 Neighbor Discovery:** The authors will send the draft to the SIXMAN mailing list for review and consider moving the document to SIXMAN entirely. * **SRv6 Egress Protection:** The authors will copy SPRING and LSR on the working group last call. * **BFD and VRRP:** The working group will proceed with one solution, and the group overwhelmingly preferred one of the solutions. * **Lightweight Routing using LLDP:** The WG raised significant concerns on the design. The authors should address these issues and propose an updated design. * **SRV6 Next Hop YANG Model:** Authors will coordinate with existing efforts in the SPRING working group. * Authors of several new proposals were asked to clearly articulate the problems these drafts were trying to solve. ## Next Steps * Authors to update drafts based on feedback received during the meeting. * Authors to send drafts to relevant working groups for review, including SIXMAN, SPRING, and LSR. * Continued discussion on the mailing list for various topics.