Markdown Version | Session Recording
Session Date/Time: 27 Jan 2026 16:00
CDNI
Summary
The CDNI Working Group held an interim session to discuss the status of several active drafts. Key discussions focused on adjusting document milestones, clarifying technical aspects and the process for the "Edge Control Metadata" and "Control Interface and Triggers" drafts, and a significant debate regarding the scope and naming of the cdn-selection object within the "Delivery Metadata" draft. Substantial updates to the "Logging Extension Updates" draft were announced, with a call for broad review.
Key Discussion Points
- Milestone Review:
- Two deliverables, "Edge Control Metadata" and "Control Interface and Triggers," initially targeted for completion by end of 2025, are now projected for 2026. These timelines need adjustment.
- Other March deliverables (e.g., "Protected Secrets," "Delivery Metadata") also require timeline review for feasibility.
- Edge Control Metadata (draft-ietf-cdni-edge-control-metadata-08):
- A key issue was the normative dependency on the "Processing Stages" document, which is not yet an RFC.
- The latest draft (version 08) successfully uncouples this dependency, allowing "Edge Control Metadata" to progress independently.
- Guillaume, as shepherd, indicated minor textual comments remain for version 08, which he will post to the mailing list.
- The chairs clarified that minor textual changes typically do not necessitate a new Working Group Last Call (WGLC).
- Control Interface and Triggers (draft-ietf-cdni-control-interface-triggers):
- Draft 18 introduced significant new material, including a refactored "trigger collection" concept to "trigger index" and detailed processing of multiple extensions.
- "Processed" Status (from RFC 8007): Debate on its compatibility with dependency and execution policies. The consensus was to clarify that triggers using "processed without acknowledgment" status cannot be reliably used for dependencies and to discourage its use in such contexts. DCDNs utilizing this status are expected to manage internal consistency.
- Offline Cache Nodes: The responsibility for managing the state of offline or rejoining cache nodes, particularly regarding previously issued triggers, was clarified to be an internal DCDN concern. Triggers on offline nodes should not count towards completion.
- Error Handling: A need was identified to consolidate error codes and their meanings into a single chapter or table for implementers.
object-list(hrefvs.data): Confusion was raised regarding the duplication ofhrefanddataoptions within theobject-liststructure, suggestinghrefcould be a degenerate case ofdata. This needs further offline discussion.- Terminology ("surrogates"): The term "surrogates" (a holdover from RFC 1007) should be replaced with clearer terms like "cache nodes" or "DCDN."
payload-typeRegistration: Confirmed thatpayload-typeextensions can be registered with IANA without requiring a full RFC process.
- Delivery Metadata (draft-ietf-cdni-delivery-metadata):
- Discussion centered on the
cdn-selectionobject. Chris's comment questioned if the name was misleading if it only concerned transport. - Guillaume explained it relates to selecting a "virtual CDN" or a "transport arrangement" (which may or may not map to distinct physical infrastructure), particularly in the context of multicast delivery.
- Alan argued that introducing
cdn-selectionas a mechanism for selecting among multiple DCDN "entities" or capabilities (beyond a monolithic DCDN) represents a fundamental architectural change akin to virtualization, requiring far more extensive architectural discussion and text than currently provided. - Sanjay offered an interpretation that
cdn-selectionallows a DCDN to advertise different infrastructure capabilities (e.g., live streaming vs. VOD vs. multicast) and for a UCDN to delegate specifically to one of these advertised capabilities. - Alan suggested renaming it to
transport-selectionto avoid implying a more significant "CDN virtualization" concept, which has precise meaning in CDNI.
- Discussion centered on the
- Logging Extension Updates (draft-ietf-cdni-logging-extension-updates):
- Ben (author) announced a new draft with substantial changes, including support for custom fields, custom record formats, and sampling.
- The document is significantly larger and different from previous versions, requiring thorough review.
- Client Access Control (draft-ietf-cdni-client-access-control):
- Chris mentioned that comments have been posted to the mailing list for review.
Decisions and Action Items
- Milestone Updates: Co-chairs to initiate a discussion on the mailing list to update the timelines for delayed deliverables (Edge Control Metadata, Control Interface and Triggers) and review the feasibility of other March deliverables.
- Edge Control Metadata: Guillaume to post minor textual comments on the mailing list. Authors (Alfonso and Glenn) to consider these comments and potentially issue a version 09. Once satisfied, the Working Group Last Call will conclude, and the shepherd report will be prepared.
- Control Interface and Triggers:
- Authors (Alan and Jay) to incorporate feedback, including clarifications on the "processed" status, DCDN responsibility for offline nodes, and a consolidated error handling section, into a new Draft 19.
- Kevin (shepherd) to review Draft 19.
- Further offline discussion to resolve the
object-list(hrefvs.data) confusion. - Replace "surrogates" with "cache nodes" or "DCDN," defining if necessary in the terminology section.
- Once Draft 19 is stable, a Working Group Last Call will be initiated, followed by the shepherd report and submission to IESG.
- Delivery Metadata: Discussion on the naming convention (
cdn-selectionvs.transport-selection) and the conceptual scope of the object to continue on the mailing list. Alan and Sanjay to post comments. - Logging Extension Updates: All working group members are strongly encouraged to review the new draft due to its substantial changes and provide comments on the mailing list.
- Client Access Control: Working group members to review Chris's comments posted on the mailing list.
- IETF 119 Agenda Request: Co-chairs to request agenda time for the "Logging Extension Updates" document at the upcoming IETF 119 meeting in March.
Next Steps
- Continued discussions and document revisions on the mailing list for all discussed drafts.
- Preparation and submission of Draft 19 for "Control Interface and Triggers."
- Working Group Last Call for "Edge Control Metadata" following the incorporation of minor comments.
- The chairs anticipate scheduling another interim meeting after the March IETF to continue progress.