Markdown Version | Transcript | Session Recording | Session Materials

Session Date/Time: 20 Jul 2026 12:00 ✎ Suggest a correction

OPSAWG - IETF 126 Meeting Minutes

Summary

The Operations and Management Area Working Group (OPSAWG) met at IETF 126. The session featured updates on active working group items, including guidelines, scheduled OAM tests, YANG data provenance, and GPON GEM IPFIX. A large number of IP Flow Information Export (IPFIX) extensions and new proposals were introduced, prompting an architectural discussion on whether specific technology-tied IPFIX elements should be adopted within OPSAWG or inside their respective technology working groups.

Key Discussion Points

1. Administrative & Status Updates

Presenter: Joe Clarke and Benoît Claise
Slides: 1. Chairs Slides
Active Draft Updates:
- draft-ietf-opsawg-rfc5706bis: The Working Group Last Call (WGLC) is currently active. The consensus-running chair, Alvaro Retana, extended the last call to give attendees more time to submit reviews and feedback during the meeting week.
- draft-ietf-opsawg-pcapng: Michael Tuxen completed the shepherd write-up for the historical PCAP document, which has been submitted to the Area Director (Mahesh Jethandani). A volunteer is still needed to shepherd the PCAP Now Generic (pcapng) draft.
- draft-ietf-opsawg-ipfix-alt-mark: Benoît Claise (serving as shepherd) noted that the YANG module, deployment guide, and IPFIX drafts are closely interconnected and must progress together. He suggested either moving the document package to the IPPM working group or coordinating a joint last call.
General WG Practice: The chairs requested that draft authors proactively seek reviews from peers (specifically other IPFIX and YANG authors) on the mailing list rather than relying solely on the chairs or secretary to find reviewers.
AIOps Side Meeting: Mahesh Jethandani advertised the upcoming AI Ops side meeting scheduled for Friday morning.

2. Export of GPON Encapsulation Mode in IPFIX

Presenter: Thomas Graf
Slides: 2. Export of GPON Encapsulation Mode in IPFIX
Draft: draft-ietf-opsawg-ipfix-gpon-gem
Updates:
- Integrated updates defining Port ID and properties for both GEM and XGEM headers.
- Addressed review comments from Chongfeng Xie.
Discussion: Comments from Paul Aitken's expert review are still outstanding. Thomas Graf confirmed these will be addressed in the next revision, after which the document will be positioned for Working Group Last Call.

3. YANG Data Model for Network Diagnosis using Scheduled Sequences of OAM Tests

Presenter: Qin Wu
Slides: 3. YANG Data Model for Network Diagnosis using Scheduled Sequences of OAM Tests
Draft: draft-ietf-opsawg-scheduling-oam-tests
Updates: Addressed feedback from Joe Clarke, Hongwei Yang, and Hans Åselius. Corrected the order-by statement, clarified execution counts, set default values, and reused standard scheduling groupings.
Discussion:
- A design issue was raised concerning whether a 32-bit Router ID is sufficient as a unique identifier (EID) for non-router elements (e.g., Layer 2 switches, firewalls).
- Jeff Haas noted that device context matters. While a Router ID works well for routing protocols, it is not sufficient or guaranteed to be unique for non-IP elements. He advised against making it the sole unique key.
- Joe Clarke noted that if a Router ID is used, the draft must provide clear guidance on what values non-router devices (like L2 switches) should populate (e.g., management IP).
- Qin Wu presented three options for importing identifier definitions (e.g., from network topology or inventory models). Joe Clarke suggested that the authors pick their preferred approach and float it on the list.
Decision: The chairs will initiate an early YANG Doctor review while the authors resolve the identifier selection.

4. AltMark and On-Path Telemetry drafts

Presenter: Giuseppe Fioccola
Slides: 4. AltMark and On-Path Telemetry drafts
Draft: draft-ietf-opsawg-ipfix-alt-mark
Updates: Addressed comments from the mailing list and Paul Aitken's expert review. The authors improved descriptions of flow aggregation/correlation, aligned naming with conventions, clarified centralized vs. node-level computation, and corrected cross-references between the YANG and IPFIX documents.
Discussion: Mahesh Jethandani and Paul Aitken noted that they had not yet seen the authors' responses to the expert review on the list. Giuseppe Fioccola replied that they have addressed the comments locally and will publish an updated version now that the IETF submission window has reopened.

5. Applying COSE Signatures for YANG Data Provenance

Presenter: Albert Lopez
Slides: 5. Applying COSE Signatures for YANG Data Provenance
Draft: draft-ietf-opsawg-yang-provenance
Updates: Added a threat model in security considerations. Coordinated with COSE experts to request pre-allocation of seed ranges. Introduced support for multi-signatures and counter-signatures to allow incremental signing over time.
Discussion:
- Rob Wilton raised a concern that generating a signature tied to a specific encoding (e.g., JSON or CBOR) means that translating the encoding mid-pipeline will invalidate the signature. He also suggested splitting CBOR strings and binary identifiers into separate fields and queried whether the draft requires on-wire normalization of YANG data.
- Per Andersson supported data normalization, particularly for data at rest, and raised a question regarding how sparse data chunks are handled.
- Mahesh Jethandani pointed out there are still some YANG Lint augment structure errors listed on the datatracker.
- Diego Lopez and Albert Lopez clarified that they intend to handle the canonicalization/normalization issues in a separate draft to prevent stalling the progress of the main provenance draft, which they consider mature.

6. YANG deVELpment PrOCEss & maintenance (VELOCE)

Presenter: Mahesh Jethandani
Slides: 6. YANG deVELpment PrOCEss & maintenance (VELOCE)
Topic: Discussing repository hosting strategies for the VELOCE experiment.
Discussion:
- The authors presented four options for hosting: (A) GitHub, (B) IETF GitLab, (C) IANA Registry, and (D) Platform-agnostic.
- Qin Wu, Italo Busi, and Jeff Haas supported using GitHub for the initial stages of the experiment due to its familiarity, ease of use, and superior pull-request diff review system.
- Diego Lopez, Rob Wilton, and Warren Kumari emphasized the necessity of long-term IETF data ownership and archiving, noting that external platforms like GitHub might eventually disappear or become restricted.
- Dhruv Dhody and Rob Wilton clarified that the IANA registry acts as a publication target rather than an active development repository.
- Anders Starrin shared that 3GPP handles this by hosting its own internal GitLab instance to maintain complete data ownership.

7. Security Operations Fundamentals and Guidance

Presenter: Kirsty Paine
Slides: 7. Security Operations Fundamentals and Guidance
Updates: Incorporated reviews from Adrian Farrel and Mark (providing guidance on tooling capabilities). Added explicit text addressing the natural tension between user privacy and security operators' requirement for network observability. Added sections targeting dynamic and large-scale networks.
Polls:
- Poll 1: Are you interested in seeing this work progress in OPSAWG? — yes: 20, no: 2, no_opinion: 23 (total: 90)
Discussion: Benoît Claise observed that the high "no opinion" count is expected since the OPSAWG room consists of network operations rather than security operations personnel. Jeff Haas suggested forming a cross-area design team and presenting the work at the Security Area Open Meeting (SAG) to build joint ownership.

8. QUIC Transport for Network Telemetry

Presenter: Benoît Claise
Slides: 8. QUIC Transport for Network Telemetry
Topic: Proposing QUIC (using both stream and datagram frames) as a secure, congestion-controlled, and authenticated transport for telemetry protocols (IPFIX, Syslog, YANG push, etc.).
Discussion:
- James Cumming asked why the draft restricts the scope to dial-out (router-initiated) connections. Benoît Claise replied this was done to keep the initial scope manageable, though dial-in is technically possible.
- Jeff Haas warned that multiplexing high-speed data like IPFIX on a single connection could run into complex stream-level and connection-level flow control bottlenecks.
- Per Andersson asked how the transport parameters would be configured in YANG push subscriptions.
- Qin Wu noted that defining a new telemetry header extension might require reviews from the Web and Transport (WIT) area.

9. Export of BIER Information in IP Flow Information Export (IPFIX)

Presenter: Yunsong Liu
Slides: 9. Export of BIER Information in IP Flow Information Export (IPFIX)
Updates: Removed the BIFT ID field based on feedback from the BIER WG. Added three new elements: Sub-Domain (SD), Set Identifier (SI), and encapsulation type.
Discussion: Benoît Claise noted that since the feedback in the BIER WG was highly positive, the authors should consider asking for adoption directly in the BIER WG. The chairs will coordinate this with the BIER chairs.

10. Export of BGP Prefix Origin Validation in IP Flow Information Export (IPFIX)

Presenter: Yunsong Liu
Slides: 10. Export of BGP Prefix Origin Validation in IP Flow Information Export (IPFIX)
Updates: Addressed comments from Jeff Haas. Modified the IE design from a bit-vector to an unsigned integer, added references to RFC 8893, and limited the scope to the local RIB (active forwarding states).
Discussion on IPFIX Document Placement:
- Paul Aitken expressed a strong preference for keeping all IPFIX-related documents inside OPSAWG to allow the designated experts to track changes in one consolidated place.
- Rob Wilton noted that IPFIX-heavy agendas can be exhausting for members not interested in IPFIX, suggesting splitting sessions or creating an IPFIX maintenance WG.
- Jeff Haas commented that technology WGs like IDR focus on routing and control plane, whereas IPFIX is about decorating observed forwarding state. Hence, these drafts do not fit cleanly in routing-specific working groups.

11. Ordered Information Element Export in IP Flow Information Export (IPFIX)

Presenter: Benoît Claise
Slides: 11. Ordered Information Element Export in IP Flow Information Export (IPFIX)
Topic: Introducing two new set IDs (4 and 5) to let exporters signal that repeated elements in a template must be processed in the exact order they are listed.
Discussion: Joe Clarke suggested splitting the draft into two components: one describing the operational/functional requirements and another specifying the actual IPFIX elements.

12. Export of ECN Information in IPFIX

Presenter: Yali Wang
Slides: 12. Export of ECN Information in IPFIX
Updates: Merged the separate IPv4 and IPv6 ECN Information Elements (IEs) into a single protocol-agnostic element to avoid excessive flow records in large-scale deployments. Added a new IE to export ECN information from MPLS Network Action (MNA) encapsulations.

13. Export of Segment Routing Policy Attributes in IP Flow Information Export

Presenter: Shunwan Zhuang
Slides: 13. Export of Segment Routing Policy Attributes in IP Flow Information Export
Topic: Defining four new IEs (head end, color, end point, and forwarding type) to map IP traffic flows directly to Segment Routing policies.

14. Export of QUIC Information in IP Flow Information Export

Presenter: Shunwan Zhuang
Slides: 14. Export of QUIC Information in IP Flow Information Export
Updates: Aligned flow definition terminology with standard IPFIX practices and integrated elements with the standardized packet discarding framework.
Polls:
- Poll 2: Is there interest in adopting this draft in OPSAWG? — yes: 13, no: 0, no_opinion: 9 (total: 91)
Decision: The chairs will initiate a formal working group adoption call on the mailing list.

15. Export of Source Address Validation (SAV) Information in IPFIX

Presenter: Lancheng Qin
Slides: 15. Export of Source Address Validation (SAV) Information in IPFIX
Updates: Changed terminology from "SAV effectiveness" to "SAV enforcement" monitoring. Refined the text to emphasize that IPFIX only reports local observations (rather than asserting a packet is definitively spoofed), and removed prescriptive implementation details regarding sequence associations.
Next Steps: The authors requested working group adoption.

16. Export of BGP VPN Information in IPFIX

Presenter: Yali Wang
Slides: 16. Export of BGP VPN Information in IPFIX
Updates: Defined a precise BGP VPN Next Hop address and added operational guidelines detailing how to prioritize VPN-specific IEs over generic BGP Next Hop IEs.
Next Steps: The chairs will coordinate with the BEST working group chairs to determine the appropriate working group for adoption.

17. Problem Statement for Network Resilience

Presenter: Jing Zhao
Slides: 17. Problem Statement for Network Resilience
Topic: Investigating gaps in traditional network reliability mechanisms when facing cascading, correlated, and configuration-driven failures. The draft defines key capability stages: anticipate, detect, content, adapt, and learn.
Discussion: Joe Clarke indicated he would review the document and send detailed technical comments to the mailing list.

Decisions and Action Items

draft-ietf-opsawg-rfc5706bis: Alvaro Retana extended the WGLC. Working group members are requested to review and post comments.
draft-ietf-opsawg-ipfix-gpon-gem: Thomas Graf to publish a new version addressing Paul Aitken's expert review comments. Once done, the chairs will initiate WGLC.
draft-ietf-opsawg-scheduling-oam-tests:
- The chairs will request an early YANG Doctor review.
- The authors must select their preferred identifier approach for non-router devices, document it, and post their proposal to the list.
draft-ietf-opsawg-ipfix-alt-mark: Giuseppe Fioccola to publish the updated version addressing Paul Aitken's comments on the list.
draft-ietf-opsawg-yang-provenance: Albert Lopez and Diego Lopez to split the YANG canonicalization/normalization work into a separate, dedicated draft. The core provenance draft will proceed once the outstanding security and YANG Lint reviews are formally signed off.
IPFIX Drafting & Adoption Placement: The chairs (along with the Area Directors) will evaluate the appropriate working group homes for all pending IPFIX proposals (specifically those tied closely to routing/multicast technologies like BIER and BGP VPN). For general IPFIX drafts, official adoption polls will be moved to the list.

Automatic IETF Minutes