Markdown Version | Session Recording

Session Date/Time: 03 Nov 2025 14:30

DMM Meeting Minutes - IETF 124

Summary

The DMM working group session at IETF 124 included updates on several working group documents, architectural discussions related to SRv6 for mobile user plane, mobility traffic steering, and a new problem statement on inter-access mobility scaling for private wireless networks. Key discussions revolved around clarifying scope, addressing technical dependencies, and the potential for new work in the group.

Key Discussion Points

• Mobility over Transport Network Slicing for 5G (draft-ietf-dmm-tn-aware-mobility)

  • The document completed its working group last call.
  • Comments received included:
    • "Mobility aware" terminology: Concern that the term is confusing. Authors clarified that the text explains the scope in terms of transparent mobility handling by 3GPP, with the UE obtaining services transparently. The chairs suggested a mailing list discussion on potentially changing the name if confusion persists.
    • Yang model dependency: A normative dependency on an individual draft (the Yang model split out) was identified. The author clarified the Yang model draft is not a normative dependency for the main document's RFC status, suggesting a clarification to reference it as an informational draft.
    • Security Considerations: Feedback indicated the section was either too short/unclear or re-explained existing RFCs. The author's approach is to add scenario-specific details and reference RFC 9540/9443 for guidance without repeating full explanations.
  • The author will address these comments as part of the last call resolution.

• SRv6 Mobile User Plane Motivation (draft-ietf-dmm-srv6-mup-motivation)

  • Updates: Co-authorship changed. The document's status was changed to an "informational reference" to better reflect its purpose as a motivation draft.
  • Content Discussion: The draft will not be extended with lawful intercept or malicious information details to keep it lean. However, a network slicing chapter describing edge computing and low-latency slice using SRv6 MUP will be added.
  • Feedback: Jeffrey Zhang and Longui Lee raised concerns about the necessity of the document and its clarity regarding SRv6-specific advantages. They argued that many concepts apply to other transport technologies (e.g., MPLS) and that the draft's title and content could be misleading by implying uniqueness to SRv6.
  • Author's Response (Jacob Horn): Acknowledged that MUP is applicable to any transport technology, but the initial intent was to focus on SRv6 as the most advanced and best fit. He committed to refining the language to generalize the scope or mention other possibilities, making it absolutely clear that the draft does not claim SRv6 as the only solution.
  • Next Steps: Authors will update the draft to clarify the scope and language. The intent is to request a working group last call at the next IETF meeting.

• SRv6 Mobile User Plane Architecture (draft-ietf-dmm-srv6-mup-architecture)

  • Updates (Satoru Satoh): Clarified the MAPC API definition, emphasizing that MAPC retrieves/receives session information and that the architecture is independent of specific mobile service systems. Specific MAPC API details are out of scope for this document.
  • Security Considerations: The author presented a plan to fill the security section with points on avoiding spoofing/tampering, adhering to RFC 7454 for ISIS/BGP, restricting BGP sessions, ensuring proper route propagation, and referencing SRv6 data plane security from RFC 8750/8996.
  • Clarification (Jeffrey Zhang): Re-emphasized that this work does not define a new architecture for 3GPP mobile user plane but rather shows how to implement it in a distributed fashion "under the hood" of existing 3GPP architecture.
  • Next Steps: The author plans to update the security considerations section and then request a working group last call.

• TNAWI Mobility Yang Model (draft-john-dmm-tn-aware-mobility-yang)

  • Status: This document was split from the main TNAWI draft and is not yet a working group document.
  • Updates (John Kaukeli Malin): No changes to the core Yang model logic. The model was recompiled, and security considerations were updated to reflect latest best practices (e.g., NETMOD RFC 8060). An appendix was added with an example configuration applying the Yang model with a Layer 3 UDP tunnel bearer.
  • Next Steps: The working group chairs will issue an adoption call for this document after IETF 124.

• Mobile Traffic Steering (draft-olafsson-dmm-mobile-traffic-steering)

  • Updates (Marco Olafsson): Version 4 was published, providing clarifying details and addressing comments received, particularly from John Kaukeli Malin.
  • Scope: Describes a standardized way for mobile communication systems to interface with transport and data networks for end-to-end path steering. It abstracts from detailed mobile systems, using a control/data plane separated framework.
  • Use Cases: Discusses proactive and reactive user plane anchor (UPA) relocation, inter-UPA communication, and ephemerality in future networks.
  • Frameworks: Proposes controller-based and decentralized deployment options.
  • Comments Addressed:
    • Value of the draft in identifying a gap in 5G for reselecting and steering flows.
    • Clearer differentiation of "traffic treatments" (focus on path steering initially; QoS is an "advanced" treatment).
    • Deployment options described earlier in the abstract/introduction.
    • Clarification of proactive/reactive UPA relocation from the transport network's viewpoint.
  • Further Discussion:
    • Terminology: John Kaukeli Malin suggested clarifying vague terms like "flexible," "dynamic," and "reconfigurability." The chair recommended eliminating such unqualified terminology in a technical specification.
    • QoS: John asked about incorporating QoS. The chair suggested keeping QoS out of scope for now to avoid merging complex issues. Authors agreed to keep it separate initially.
    • 3GPP Alignment: Discussion reiterated that the draft aims to be aligned and applicable to 3GPP architectures (e.g., 5G+, 6G) without modifying existing 3GPP interfaces or specifications. The C1/C2 interfaces are abstracted to be applicable to future evolutions.
    • Added Value (Lionel Morand): Raised questions about the real added value for existing vs. new deployments (6G). It was agreed the document is more suited for next-generation systems where 3GPP is more flexible to adopt new approaches.
  • Next Steps: Authors plan to extend operations sections, add clarifications, unify the information model format in the appendix. Carlos Benados's review (received offline) will be posted to the mailing list. The working group chairs will issue an adoption call after IETF 124.

• Inter-Access Mobility Problem Statement (draft-grayson-dmm-inter-access-mobility-ps)

  • New Topic (Mark Grayson): Introduced a new problem statement concerning the scaling of inter-access mobility and roaming between networks, particularly for small, private wireless network deployments (Wi-Fi and private cellular).
  • Background: Anticipates millions of private cellular networks by 2030. Current cellular roaming solutions (GSMA IR.21, IPX) don't scale to this. 3GPP SA1 introduced SNPN cellular hotspots, mimicking Wi-Fi hotspots, but a scaling challenge exists with potentially 1000x more networks and signaling.
  • Challenge: Different roaming signaling protocols (Radius, Diameter S6a, HTTP/2 N32). Server-initiated signaling (e.g., for credit updates, user relocation) is difficult across enterprise firewalls, which typically block unsolicited inbound connections.
  • Existing Work: Radext WG's "Reverse COA" draft addresses this for Radius by sending server-initiated messages down an established outbound TLS tunnel (RADSEC).
  • Problem Statement:
    • Radext is progressing for Radius-based networks.
    • 3GPP SA3 is studying N32 enhancements for 5G (SNPN) to permit server-initiated signaling using the same outbound socket (similar to Reverse COA).
    • Gap: No standardized approach exists for Diameter (LTE, predominant in private cellular for next 5 years) to handle server-initiated messages through NAT/firewalls.
  • Question to WG: Is the scaling of roaming and mobility signaling for small, localized, distributed private wireless networks within the DMM charter? If so, potential outputs could be recommendations on adaptations, interworking, or solutions for N32/S6a perimeter firewall traversals.
  • Discussion:
    • Scope: Documenting the problem itself was seen as valuable, given its cross-entity nature (IETF, 3GPP).
    • Identity Providers: Tianji Jiang highlighted the difference in scale and nature of identity providers between cellular (SIM-based) and Wi-Fi. Mark Grayson noted that cellular is evolving with ESIMs to support multiple credentials, bringing similar characteristics.
    • Administrative Domains: Clarified focus on inter-administrative domain roaming, not intra-domain mobility.
    • Potential Solutions: Lionel Morand suggested that if 3GPP is developing a solution for SNPN based on N32, it might be out of DMM scope. He expressed doubt about new Diameter work but saw value in IETF providing recommendations to 3GPP on interworking with Radext solutions for SNPNs. He suggested focusing efforts on how Radext/Reverse COA could be applied to SNPN roaming.
    • Chairs' View: Documenting the problem can be very useful, potentially leading to recommendations for 3GPP, similar to past IETF work on IPv6 support.

Decisions and Action Items

• draft-ietf-dmm-srv6-mup-motivation:
  • Decision: Authors will refine the language to clarify that its concepts are not unique to SRv6 but applicable to other transport technologies, while maintaining a focus on SRv6.
  • Action Item: Authors to submit a revised draft with these clarifications and prepare for a WG last call at IETF 125.
• draft-ietf-dmm-srv6-mup-architecture:
  • Decision: The security considerations section needs to be comprehensively updated.
  • Action Item: Authors to update the security considerations section and then request a WG last call.
• draft-ietf-dmm-tn-aware-mobility:
  • Action Item: Authors to address the comments received during the working group last call, including clarifications on "mobility aware" terminology, the normative dependency on the Yang model, and the security considerations section.
• draft-john-dmm-tn-aware-mobility-yang:
  • Action Item: WG Chairs will issue an adoption call for this draft after IETF 124.
• draft-olafsson-dmm-mobile-traffic-steering:
  • Decision: Vague terminology (e.g., "flexible") should be clarified or removed. QoS will remain out of scope for now.
  • Action Item: Authors to refine vague terminology, extend operations sections, add clarifications, and unify the format of the information model in the appendix. Carlos Benados is requested to post his review comments to the DMM mailing list.
  • Action Item: WG Chairs will issue an adoption call for this draft after IETF 124.
• draft-grayson-dmm-inter-access-mobility-ps:
  • Action Item: Authors to continue gathering feedback from the WG on whether this problem statement fits within the DMM charter and how to proceed (e.g., as recommendations to 3GPP, focusing on Radext interworking, or specific solutions).

Next Steps

• WG Chairs to issue adoption calls for draft-john-dmm-tn-aware-mobility-yang and draft-olafsson-dmm-mobile-traffic-steering after IETF 124.
• Authors of draft-ietf-dmm-srv6-mup-motivation and draft-ietf-dmm-srv6-mup-architecture to prepare for WG last calls at IETF 125.
• Continued discussion on the draft-grayson-dmm-inter-access-mobility-ps problem statement to determine working group scope and next steps.