Session Date/Time: 05 Nov 2025 14:30

IPPM Session Minutes

Summary

The IPPM and BMWG Working Groups held a joint session to discuss ongoing document progress, propose new work, and explore the potential for consolidation of the two working groups. Several working group documents from both IPPM and BMWG were presented, detailing updates, addressing review comments, and indicating readiness for Working Group Last Call or IESG publication. A significant portion of the session was dedicated to a presentation on the rationale and proposed plan for consolidating IPPM and BMWG, which generated lively discussion with both support and strong opposition, leading to a decision to continue the discussion offline and at an upcoming interim meeting.

Key Discussion Points

Welcome and Joint Meeting Rationale: The chairs welcomed participants, outlined IPR and conduct guidelines, and explained the rationale for the joint IPPM/BMWG meeting: adjacent topics warranting cross-discussion and exploration of closer collaboration, including a potential joint charter. Chin presented a detailed proposal for this at the end of the session.
BMWG Charter Overview: Giuseppe described the BMWG charter, focusing on producing recommendations for performance characteristics, KPIs, and benchmarks for network devices, systems, and services in a controlled lab environment.
IPPM Charter Overview: The chair outlined the IPPM charter, emphasizing its focus on producing measurements and protocols for operational networks, particularly internet data delivery services and applications over transport layer protocols. The distinction between lab (BMWG) and operational (IPPM) environments was highlighted.
Question on Consolidation (Mathieu Konstantanovich): Mathieu Konstantanovich (FDI open source/Cisco) inquired if the discussion about a joint charter implies a merger of both WGs or a looser coupling. The chair deferred a detailed answer to Chin's presentation at the end of the session.
BMWG Document Updates:
- draft-ietf-bmwg-mlr-search submitted to IESG for publication.
- draft-ietf-bmwg-yang-network-tester passed WGLC; new revision in preparation to address YANG doctors' review; cross-posting to unmop and Ops-RG.
- draft-ietf-bmwg-sr-benchmarking preparing for WGLC.
- draft-ietf-bmwg-power-benchmarking adopted this week.
- New milestone under AD review.
- Future discussion items: Revisiting RFC 2544 and managing expired WG documents.
IPPM Document Updates:
- draft-ietf-ippm-capacity-prod (Capacity product) finally approved by IESG after resolving security concerns. Rudiger was thanked for his significant work.
- draft-ietf-ippm-asym-pkts (Asymmetrical packets) and draft-ietf-ippm-hybrid-two-step passed to IESG, with the latter receiving Ops-AD review.
- draft-ietf-ippm-quality-outcome passed WGLC and shepherd review, submitted to IESG.
- draft-ietf-ippm-on-path-active passed WG adoption call.
- draft-ietf-ippm-integrity-yang adoption call is running; WG urged to review and provide input.
YANG Data Model for Network Tester Management (Vladimir):
- Presented draft-ietf-bmwg-yang-network-tester.
- Addressed Andy Berman's YANG doctors' review comments, fixing ~60% of issues.
- Key unresolved issue: choice between binary (base64) and hex-string for specifying packet frame data in YANG models, balancing usability for test case writers (preferring hex) against tool encoding (base64 for NetConf). Discussion may continue in NetMod WG.
- New 07 version tested in hackathon; awaiting N-OPSA group reviews.
On-Path Telemetry for Active Measurement (Giuseppe):
- Presented draft-ietf-ippm-on-path-active-measurement.
- Scope: Hop-by-hop and end-to-end active measurements using active test packets augmented with hybrid methods (e.g., ICMPv6, O-WAMP, STAMP) and IPv6 hop-by-hop options (Altmark, IOM) or MPLS data plane.
- Key update: Included MNA sub-stacks for IOM and Altmark, added RFC 9714 example for MPLS encapsulation for Altmark, and reference to SRv6 Altmark application.
- Document is stable; comments welcome for review.
SR Benchmarking Methodology (Louis):
- Presented draft-ietf-bmwg-sr-benchmarking.
- Scope: Defines standard methodology for benchmarking Segment Routing (SR-MPLS, SRv6, including compression).
- Update: Added new Section 7 on scaling tests, defining baseline and various scalability dimensions (seed scale per segment list, segment list scale, candidate paths, SR policies). New report parameters also added.
- Next steps: Address remaining comments, plan for WGLC, evaluate liaison from ITU-T SG11.
STAMP Extensions for Reflecting IP Headers (Rakesh Gandhi):
- Presented draft-ietf-ippm-stamp-ip-header-reflection.
- Requirements: Hop-by-hop and end-to-end (one-way and two-way) measurements leveraging existing midpoint implementations for STAMP and IPv6 extensions.
- Updates: Draft split into IP and MPLS data planes. Clarifications and elaboration of processing based on William's open-source implementation feedback. IANA review comments addressed.
- Ready for WGLC, pending a few recent review comments.
STAMP Extensions for Reflecting MPLS Headers (Rakesh Gandhi):
- Presented draft-ietf-ippm-stamp-mpls-header-reflection.
- Requirements: Similar to IP version but for MPLS data plane, incorporating MNA working group documents.
- Updates: Draft split, clarified procedures to align with latest MNA post-stack header draft, IANA review comments addressed.
- Greg Murski (Erickson) raised concerns about how the reflector would receive MPLS information if MPLS encapsulation is removed before reaching the application layer, potentially invalidating hop-by-hop or S-2 measurement for MPLS. The chair suggested taking this fundamental discussion offline to resolve.
Benchmarking in Containerized Network Infrastructure (Mingok):
- Presented draft-ietf-bmwg-container-infra-benchmarking.
- Scope: Focuses on container architectural and CNI aspects affecting benchmarking results (networking models, resource configurations).
- Update: Aligned scope with a similar individual draft (CNI Telco Cloud Benchmarking) – WG document focuses on general container infrastructure, while the individual draft focuses on Telco cloud CNI specifics. Added a new subsection on container deployment environment configuration parameters.
- Believed ready for WGLC in December (though Giuseppe noted potential milestone delays).
IOM Data Integrity (Justin):
- Presented draft-ietf-iom-data-integrity.
- Updates: Addressed Giuseppe's review (clarified method description) and Ben's security directorate review.
- Resolved nonce reuse issue by clarifying that nodes must not reuse the same nonce with the same key, even after crashes/reboots. Key exchange/rotation out of scope.
- Addressed interop issue by proposing an IANA registry for header fields that require integrity protection with masks for each field, as IOM option types lack versioning.
- Believed ready for a second WGLC. Chair urged WG members to read and review.
Responsiveness Control (Stuart Cheshire):
- Presented draft-ietf-ippm-responsiveness.
- Document is well-polished and reviewed, but recent real-world testing (Kunder Sherpa at Nokia on L4S) revealed issues.
- Three problems identified:
  1. Bug in Accurate ECN specification's heuristic for detecting broken middleboxes, unnecessarily disabling ECN in the presence of reordering.
  2. RPM test's use of multiple parallel connections interacts poorly with traffic policers that use ECN marking, leading to excessive marking and packet drops when new connections ramp up faster than existing ones reduce rate.
  3. Source buffering delay: Default 16KB backlog can add significant latency on low-speed connections (e.g., 10Mbps), skewing results.
- Proposed to keep the document as a "living document" for continued experimentation and revision before publication, emphasizing quality over rushed RFC. The chairs expressed support for this approach.
Benchmarking Methodology for Source Address Validation (Ivan Liu):
- Presented draft-liu-bmwg-sav-bench.
- Scope: Defines methodologies for benchmarking performance of intradomain and interdomain Source Address Validation (SAV).
- Updates: Addressed comments regarding test setup variations (spoofed traffic types, ratios), clarified performance indicators (false positive/negative, convergence time, forwarding rate), added considerations for standard deviation and varying SAV table sizes, clarified optionality of control plane tests for non-protocol-specific SAV, updated terminology, and added new testing scenarios (FRR, policy-based routing, partial deployment).
- Requested WG adoption. Giuseppe indicated readiness for an adoption call.
STAMP Class of Service Extension (Greg White):
- Presented draft-white-stamp-cos-ext.
- Proposed update to STAMP Class of Service TLV to enable bi-directional ECN traversal and performance measurement.
- History: Discussions at IETF 122 (new TOV, revising existing TOV, ambiguity resolution by Jaumin, IETF guidance on ECN by Gorry Fairhurst).
- Updates: Added two new fields to reserved bits in existing CoS TLV: desired return path ECN value (IPECN) and a Return Path ECN (RPE) flag.
- Ambiguity for RPE flag (if reflector implements new TOV but cannot set desired IPECN value or if sender wants 00 ECN) proposed to be resolved with "Option 3" (upper bit of RPE indicates new reflector, lower bit indicates successful IPECN setting).
- Implementation in Tea Party (Will Hawkins) and Cable Labs (non-public).
- Rakesh Gandhi inquired about interaction with IP header reflection draft (which would also reflect DSCP/ECN) and the possibility of configuring ECN on the reflector. Greg Murski highlighted the advantage of dynamic per-packet testing over configuration for network behavior monitoring.
- Document is a candidate for adoption call.
STAMP Extensions for Measuring Residual Bit Error Rate (Rakesh Gandhi):
- Presented draft-ietf-ippm-stamp-ber-ext.
- Objective: Measure residual bit error rate (BER) that escapes CRC/FEC and affects UDP checksum, complementing existing packet loss measurements.
- Protocol extensions: Uses existing extra padding TLV, defines two new TLVs for bit pattern (optional, with default) and bit error count. Reflector compares pattern, counts errors, sends back.
- Use cases: Measuring link BER, configuring padding size/pattern/interval.
- Implementation: William implemented in open source (Tea Party), providing feedback.
- Requested WG adoption. Chairs noted good discussion and support, confirming it's on the queue for adoption.
Potential Consolidation of IPPM and BMWG (Chin):
- Presented a detailed rationale and proposal for consolidating IPPM and BMWG into a new working group, BIPP (Benchmark IP Performance Amendment).
- Rationale: Both WGs operate independently with limited cross-coordination, have overlapping interests/technologies, BMWG has seen declining participation, IPPM has slow standardization.
- Benefits: Single point of contact for SDO collaboration, common building blocks (security framework, applicability of metrics), increased contributor/reviewer base.
- Charter Comparison: Acknowledged IPPM's focus on operational networks vs. BMWG's on lab environments, but highlighted cross-border work in both.
- Common Ground: Proposed consolidation built on common ground of performance metric directory, single SDO contact, and common framework/guidelines.
- Open Issues: Common framework design, leveraging performance metric directorate, dispatcher role for technology-specific performance management, and cross-SDO collaboration.
- Proposed Timeline: Continue discussion on WGs mailing lists, interim meeting (Nov 20th) to discuss draft consolidated charter, aim for IESG review by end of 2026.
- Discussion:
  - Greg Mirsky (Erickson) strongly opposed the consolidation, emphasizing the fundamental difference between lab benchmarking (reproducible, controlled) and operational network measurement (current condition, no reproducibility expectation). He argued there's no synergy, questioned the perceived benefits, and expressed concern about "land grab" and lack of community feedback on the idea.
  - Mathieu Konstantanovich (FDI open source/Cisco) raised concerns about diluting participation and flooding the new WG with diverse topics, as different people are interested in IPPM vs. BMWG work.
  - Another attendee questioned whether the scope of the consolidated WG would cover both original charters fully or only their intersection. Chin clarified the intent is to cover both, finding common ground while allowing specific work.
  - An attendee expressed positive sentiment, seeing the combined scope covering "pre-production" testing that other SDOs are exploring.
  - A final quick comment reiterated the preference to keep scopes separate due to fundamental differences and the large amount of work involved in merging and revisiting RFCs.

Decisions and Action Items

Decision: draft-ietf-ippm-capacity-prod (IPPM Capacity Product) has been approved by IESG.
Decision: draft-ietf-bmwg-power-benchmarking (BMWG Power Benchmarking) has been adopted as a working group document.
Action Item: Giuseppe to send an email to the BMWG mailing list to solicit feedback on expired working group documents to determine if there's continued interest or if they should be moved to an expired state.
Action Item: The IPPM chairs will consider progressing draft-ietf-iom-data-integrity to a Working Group Last Call and strongly urge working group members to review the latest version due to its importance and complexity.
Action Item: Stuart Cheshire will propose the removal of the ill-advised heuristic in the Accurate ECN specification within the TCPM working group.
Action Item: Stuart Cheshire and Kunder Sherpa will work offline to tune the RPM code to better handle parallel connections and interactions with traffic policers.
Action Item: Stuart Cheshire will investigate and tune the source buffering size in the RPM test implementation to minimize self-induced latency, especially on low-speed links.
Action Item: The BMWG chairs will consider running an adoption call soon for draft-liu-bmwg-sav-bench (Benchmarking Methodology for Source Address Validation).
Action Item: The IPPM chairs will consider draft-white-stamp-cos-ext (STAMP Class of Service Extension) as a candidate for an upcoming adoption call.
Action Item: Greg Murski and Rakesh Gandhi are requested to continue their discussion offline or on the mailing list regarding the behavior of STAMP reflectors and MPLS encapsulation removal for draft-ietf-ippm-stamp-mpls-header-reflection.
Action Item: The IPPM and BMWG chairs will facilitate continued discussion on the mailing lists regarding the potential consolidation of the two working groups, noting the strong feedback received. This topic will be revisited at the joint interim meeting on November 20th.

Next Steps

Working group members are urged to review and provide feedback on documents nearing Working Group Last Call or adoption.
Authors are to continue addressing review comments and preparing documents for the next stages of publication.
Further real-world testing and refinement of the Responsiveness Control (RPM) draft will continue, with the document remaining "living" for now.
The discussion on the potential consolidation of IPPM and BMWG will be continued on the respective mailing lists and at the joint interim meeting scheduled for November 20th.