**Session Date/Time:** 01 Jul 2022 15:00

# [RAW](../wg/raw.html)

## Summary

This interim meeting of the RAW Working Group aimed to canvas the working group for open issues and discuss the current state of the architecture draft before IETF 114. Key discussions revolved around the scope and content of the architecture document, particularly its relationship with DetNet, the level of detail for OAM, and the inclusion of specific use cases. Given the absence of key co-authors, the working group decided to postpone the Working Group Last Call (WGLC) until after IETF 114 to allow for full input and review.

## Key Discussion Points

*   **Meeting Purpose:** The meeting was convened to solicit feedback and identify issues with the RAW architecture draft, which is considered fairly mature, in preparation for IETF 114.
*   **Karina's XML/RFC Translation Issue:** Karina raised a problem with converting her XML draft to RFC format and requested help from the mailing list.
*   **Multi-Domain RAW:** Carlos invited working group members interested in his multi-domain RAW draft to contact him for potential collaboration.
*   **Architecture Document Scope:** Pascal initiated a discussion on whether the architecture document sufficiently covers the RAW architecture in general or focuses too narrowly on the Packet Selection Controller (PSC), and what the overall expectations are for RAW.
    *   The absence of co-authors Lou and Rick made direct discussion on specific proposed architecture sections challenging.
*   **Inclusion of Examples/Details:** Carlos suggested adding more detailed examples of the architecture, possibly in an annex, and clarifying the roles of distributed vs. centralized PSC.
    *   Pascal expressed concern that including specific examples in the architecture document could inadvertently constrain future thinking about RAW's applicability. He suggested such "how-to" details might be more appropriate for the *framework* document, which describes the implementation.
*   **RAW and DetNet Interoperability/Relationship:**
    *   Carlos sought clarification on how RAW is envisioned to integrate with DetNet, especially in mixed wired/wireless multi-domain scenarios.
    *   Janos emphasized that RAW should ideally function as a *wireless extension* of DetNet, rather than requiring explicit interworking mechanisms. He suggested that if RAW is well-defined, it should seamlessly complement DetNet.
    *   Pascal concurred, stressing that RAW must be fully compatible with DetNet. He noted that while RAW is particularly crucial for wireless due to its complexity, its PSC mechanisms could also apply to wired interfaces.
    *   Pascal outlined two primary RAW use cases concerning DetNet:
        1.  **RAW "over" DetNet (Mesh Case):** RAW provides new services on top of an existing DetNet-supported network, offering end-to-end determinism within a mesh.
        2.  **RAW "preceding" DetNet (Access Case):** RAW manages the access link (e.g., 5G/Wi-Fi interfaces on a device) for high reliability, and then DetNet takes over for the core network. In this case, RAW cannot guarantee end-to-end deterministic latency (as DetNet requires full end-to-end control) but focuses on delivery optimization at the access. This "access use case" was highlighted as a significant area of interest for many stakeholders.
    *   Greg viewed RAW and DetNet as complementary. He suggested RAW could provide access (e.g., non-3GPP) that then connects to a 5G network slice where DetNet provides transport network connectivity, avoiding direct replacement or strict "end-to-end" terminology for 5G contexts.
*   **OAM in RAW Architecture:**
    *   Greg raised questions about OAM metrics, additive metrics, and the disaggregation of service level objectives within the architecture.
    *   Pascal questioned whether the architecture document provides sufficient detail on OAM, including responsibilities, components, and potential diagrams, without over-specifying.
    *   Greg, as an OAM expert, stated that the current RAW architecture document *sufficiently covers OAM* for its current stage. He suggested that more detailed OAM aspects specific to RAW or particular use cases could be addressed in separate, complementary documents, similar to how DetNet has its own OAM framework documents.

## Decisions and Action Items

*   **Pascal:** Will incorporate the discussion points, particularly diagrams illustrating the functional relationship between RAW and DetNet (RAW over DetNet vs. RAW preceding DetNet for access), into the architecture document.
*   **Karina:** Will post an updated version of her draft next week, reflecting addressed comments.
*   **Working Group Last Call (WGLC):** The WGLC for the architecture document will be delayed until *after* IETF 114, likely in late August or September, to allow for more input (especially from Lou) and avoid timing conflicts with IETF 114 and summer vacations.

## Next Steps

*   **Next Interim Meeting (July 15th):** This meeting will be an opportunity to gather further inputs, specifically from Lou if he returns from vacation, on additional topics or changes needed before WGLC.
*   **Document Updates:** Pascal will update the architecture draft to reflect the discussed architectural relationships. Karina will provide her updated draft.
*   **IETF 114 Presentation:** One of Karina's co-authors will present brief updates based on feedback at IETF 114 in Philadelphia.
*   **Continued Review:** The working group will continue to review the architecture document for readiness for WGLC, ensuring it strikes the right balance of detail for OAM and clear definition of the DetNet-RAW relationship.