Session Date/Time: 24 Jul 2025 15:00

spring

Summary

The SPRING working group meeting covered updates on several drafts, including SRv6 security considerations, performance measurement using STEM for SR networks, SRv6 YANG models, SR policy list optimization, eligibility, and multi-path traffic engineering for segment routing. Key discussions revolved around security, YANG modeling, data plane considerations, and the applicability of various techniques in different network scenarios. Several drafts are nearing working group last call.

Key Discussion Points

SRv6 Security Considerations: Clarification needed on the scope of control and management planes. Consideration of PCEP and BGP as attack vectors.
Performance Measurement using STEM for SR Networks: Discussion on splitting the draft into two documents for SRMPLS and SRv6. Concerns raised regarding dependencies on MNA and potential alternatives using native SRMPLS. Agreement that MPLS encoding questions should be addressed in the MPLS working group.
SRv6 YANG Models: Discussion on aligning with RFC 9800 (Compression). Debate on whether compression should be a feature or a base requirement in the YANG model. Feedback to use IANA-maintained modules for SRv6 endpoint types.
SR Policy List Optimization: Concerns regarding the structure of the draft and the use of normative language for PCEP and BGP. Need to present the concept in a generic, informational way.
Eligibility: Discussion about the impact of multiple eligible candidate paths (CPs) on device scalability and control plane complexity.
Cid as Source Address SRv6: Debate on the validity of the proposal in a trusted domain. Concerns regarding the potential for source spoofing. Clarification that the main goal is to solve a deployment problem related to stateful firewalls. Discussion about scenarios where a VPN advertises multiple SIDs.
176 path verification: Recursive ALG method is proposed to improve the accuracy of path verification by including the verification order
Multi-Path Traffic Engineering for Segment Routing: Presentation on using junction segments and binding SIDs to create directed acyclic graphs (DAGs) for traffic engineering. Concerns about using the color field for DAG identification and local optimization. Zaffirle claims that zero change is needed and transit policies as are already covers the whole thing because of the ECNPAW and CNB concept. It was noted that MBTE is already supported in SR and more importantly, by SRMPLS and SRV6.

Decisions and Action Items

SRv6 Security Considerations: Volunteer found for shepherd role.
Performance Measurement using STEM for SR Networks: Authors will consider splitting the draft and consult with the MPLS working group regarding MNA dependencies.
SRv6 YANG Models: Authors will align with RFC 9800 and explore using IANA-maintained modules.
SR Policy List Optimization: Authors will revise the draft to present the concept in a generic, informational way, avoiding normative language for specific protocols.
Eligibility: Further discussion on the mailing list regarding scalability considerations.
Cid as Source Address SRv6: Further discussion on the mailing list to address concerns about the validity of the proposal and the potential for source spoofing.
176 path verification: Authors will update the code to TBD for TRB code
Multi-Path Traffic Engineering for Segment Routing: Explore weight zero support in estep policies.

Next Steps

Continue discussions on the mailing list for drafts where open issues remain.
Schedule interim meetings focusing on SR and NRP security topics.
Authors to address outstanding comments and revise drafts accordingly.
Working group to consider adopting drafts nearing completion.
Request early reviews from YANG doctors for YANG models.