Markdown Version | Session Recording | Session Materials
Session Date/Time: 18 Mar 2026 01:00
NETMOD
Summary
The NETMOD Working Group met at IETF 125 to discuss the status of active drafts, provide updates on YANG versioning and packages, and establish a baseline for YANG-next (YANG 2.0). Key achievements since the last meeting include the publication of RFC 9899, RFC 9911, and RFC 9922. The session focused on refining guidance for IANA and the RFC Editor regarding YANG modules, advancing the YANG Packages and Schema Comparison drafts, and initiating the formal process for YANG-next by introducing a protocol-neutral baseline.
Key Discussion Points
Session Intro & WG Status
- Presenters: Kent Watsen, Lou Berger
- Slides: 01) Session Intro & WG Status
- Status: Three new RFCs published (9899, 9911, 9922). Several documents are in the RFC Editor or IESG queues.
- Versioning Cluster: Discussion regarding draft-ietf-netmod-yang-module-versioning, draft-ietf-netmod-yang-semver, and draft-ietf-netmod-yang-filename-extension. Per Andersson noted updates to IANA considerations and ongoing discussion regarding filename separators (e.g., using something other than
#). Joe Clarke noted that draft-ietf-netmod-yang-semver will receive an update to address a typo and feedback from Tony Lee.
Guidance for Managing YANG Modules in RFCs and IANA Registries
- Presenter: Rob Wilton
- Draft: [draft-ietf-netmod-iana-yang-guidance]
- Slides: 02) Guidance for Managing YANG Modules in RFCs and IANA Registries
- Discussion:
- The draft provides guidance to IANA and the RFC Editor on processing YANG modules, specifically regarding semantic versioning.
- IANA Registry Columns: Discussion on whether to list modules by revision date, version, or both. Amanda Baber (IANA) confirmed IANA can support both. The group agreed that the registry should show both revision date and version number to facilitate transition.
- Errata: Rob Wilton proposed deferring the handling of errata (incorporating technical fixes into new module versions without a full RFC bis) to the "Velocity" effort. The group agreed to keep errata out of scope for the current draft.
- 2119 Language: The group agreed that using RFC 2119 keywords is appropriate for this Informational document to provide clear direction to IANA.
YANG Packages
- Presenter: Rob Wilton
- Draft: [draft-ietf-netmod-yang-packages]
- Slides: 03) YANG Packages
- Discussion: Recent updates include improvements to terminology, circular import rules, and handling redundant information. Significant work was done on "mount points," adopting a simplified solution where child packages are unioned by default, with a "clean slate" flag available to start from scratch. Rob Wilton plans to have the document ready for Working Group Last Call (WGLC) before IETF 126.
YANG Schema Comparison
- Presenter: Per Andersson
- Draft: [draft-ietf-netmod-yang-schema-comparison]
- Slides: 04) YANG Schema Comparison
- Discussion: The reference implementation in
libyangis now complete, supporting both compiled and parsed schema changes. Per Andersson noted that the module is stable and the focus is now on gathering user feedback and adding examples. Joe Clarke suggested referencing the versioning drafts for terminology like "backwards compatible."
MAC Address as String
- Presenter: Scott Mansfield
- Slides: 05) MAC Address as String
- Discussion: The problem stems from differing canonical formats between IEEE (dashes/uppercase) and IETF (colons/lowercase). Scott Mansfield suggested a generic solution in YANG (e.g., a "canonical" flag or display hints) to allow programmatic equivalence checks. Rob Wilton suggested this be discussed within the YANG-next effort.
YANG Configuration Templates
- Presenter: Shufan
- Draft: [draft-ietf-netmod-yang-configuration-templates]
- Slides: 06) YANG Configuration Templates
- Discussion:
- Validity: Does a template need to be valid at definition time? The group reached a consensus on "best-effort" validation: servers should validate what is possible (e.g., syntax, ranges) but cannot enforce constraints that depend on the final expansion context (e.g.,
mandatory trueormin-elements). Full validation occurs at theintendeddatastore after expansion. - Modeling Content: The authors favor using
anydatato model template content. However, this relies on progress of draft-ietf-netmod-yang-anydata-validation. A poll indicated the room supportsanydataprovided the validation draft moves forward.
- Validity: Does a template need to be valid at definition time? The group reached a consensus on "best-effort" validation: servers should validate what is possible (e.g., syntax, ranges) but cannot enforce constraints that depend on the final expansion context (e.g.,
XML Encoding of Data Modeled with YANG
- Presenter: Kent Watsen
- Slides: 07) XML Encoding of Data Modeled with YANG
- Discussion: This work extracts XML-specific encoding rules from RFC 7950 into a standalone document, making YANG protocol-neutral. A poll showed high interest in the document. The group preferred waiting to publish it alongside the YANG-next specification rather than as a standalone update to 1.1.
A Baseline for YANG-next
- Presenter: Kent Watsen
- Slides: 08) A Baseline for YANG-next
- Discussion:
- YANG 2.0: The consensus from interim meetings is to pursue YANG 2.0 as a minimally non-backwards compatible version of 1.1.
- Baseline: The baseline draft removes XML and NETCONF specificities from the core language specification.
- Workflow: Kent Watsen proposed a GitHub-centric workflow for merging changes into the baseline. Participants (Joe Clarke, Italo Busi, Scott Mansfield) generally supported the use of GitHub but emphasized keeping the process inclusive for those not using the tool and ensuring the mailing list remains the place for final consensus.
Decisions and Action Items
- Decision: The IANA registry for YANG modules will be updated to include both the revision date and the semantic version number.
- Decision: The YANG Configuration Templates draft will adopt a "best-effort" validation approach for templates at definition time, with full validation enforced upon expansion into the intended datastore.
- Action Item (Chairs): Initiate a formal adoption call for the "XML Encoding of Data Modeled with YANG" draft.
- Action Item (Chairs): Initiate a formal adoption call for the "YANG 2.0" (YANG-next baseline) draft.
Next Steps
- YANA Guidance: Rob Wilton to update draft-ietf-netmod-iana-yang-guidance to version -02 to address discussed issues, then move toward WGLC.
- YANG Packages: Authors to finalize mount point text and examples for draft-ietf-netmod-yang-packages.
- YANG-next: Once adopted, Kent Watsen will begin managing parallel branches for specific features (e.g., removing
anyxml) according to the proposed GitHub workflow.
Related Documents
draft-ietf-netmod-iana-yang-guidance, draft-ietf-netmod-yang-anydata-validation, draft-ietf-netmod-yang-configuration-templates, draft-ietf-netmod-yang-filename-extension, draft-ietf-netmod-yang-module-versioning, draft-ietf-netmod-yang-packages, draft-ietf-netmod-yang-schema-comparison, draft-ietf-netmod-yang-semver