MANET

Summary

The MANET Working Group met at IETF 126. The session covered working group document status updates, including DLEP extensions, OLSRv2, and internetworking gap analysis. Three presentations were delivered on MANET internetworking, adapting the Babel routing protocol for IEEE 802.11 mesh networks, and a proposed MANET awareness data model. Key discussions centered on the architectural assumptions of the newly adopted internetworking draft, Layer 2 routing challenges with Babel on Wi-Fi, and how to avoid duplicating existing IETF YANG models for network topology.

Key Discussion Points

1. Document Status and Administration

The Chairs presented the MANET Chair Slides July 2026 and summarized the status of several documents:

DLEP Extensions: The three DLEP extension drafts (draft-ietf-manet-dlep-radio-quality, draft-ietf-manet-dlep-radio-band, and draft-ietf-manet-dlep-channel-utilization) completed Working Group Last Call (WGLC) last year. The Chairs apologized for the delay in processing and committed to resolving outstanding items before the end of the week.
Newly Adopted Drafts:
- draft-ietf-manet-inet-gap-analysis (presented by Fred Templin).
- A draft by Christopher Dearlove was officially adopted. This short document will be sent for an early Routing Directorate review before moving to WGLC.
Individual & Older Drafts: Charlie Perkins' AODVv2 draft has had little progress since last year; next steps will be determined on the mailing list.
DLEP RFC 8175 Errata: An errata report was officially filed by Donald Eastlake (based on an issue found by Henning Rogge). Rick Taylor was asked to review the errata on the mailing list as an active author.

2. MANET Internetworking: Problem Statement and Gap Analysis

Fred Templin presented the MANET Internetworking slide deck, focusing on draft-ietf-manet-inet-gap-analysis.

Identified Gaps:
1. Multiple Heterogeneous MANET Interfaces: Supporting concurrent Wi-Fi and 5G sidelink interfaces on devices like cell phones.
2. Maximum Packet Size Diversity: Resolving MTU differences across heterogeneous media to prevent black-holing.
3. Header Compression: Mitigating overhead for small payloads.
4. MANET Border Router Discovery and Engagement: Methods for ordinary nodes to select and route packets through border routers.
5. MANET Internetworking Security: Securing client authorization and service invocation. Stuart Card noted in the chat that there is also a need to assess trust in the border routers themselves.
Discussion:
- Protocol Enumeration: Christopher Dearlove recommended listing only IETF-standardized or experimental routing protocols (such as OSPF, OLSRv2, and Babel) in the draft, omitting non-IETF protocols like Batman or AODV. Steve supported specifying concrete protocols to address real-world interoperation issues.
- Problem Statement Concerns: Lou Berger raised significant objections to the core assumptions of the draft, noting that his previously raised adoption comments regarding the presumption of a global MANET overlay and the definition of internetworking had not been resolved. He emphasized that the working group does not yet have consensus on the problem statement and should not proceed to the gap analysis phase until this is addressed.
- Platform & Scalability Assumptions: Henning Rogge questioned the practical feasibility of using smartphones as MANET nodes due to OS-level Wi-Fi limitations. He also disputed the scalability of peer-to-peer MANETs supporting billions of nodes using small host routes, suggesting instead that standard aggregated prefixes should be used.
- Alternative Approaches: Roland Bless suggested that draft-ietf-manet-olsrv2-responsive might address some of the scalability and connectivity issues highlighted in the draft, though he noted that real-world deployment experience with it is still limited.
- Next Steps: Donald Eastlake suggested holding an ad hoc or interim meeting to resolve the foundational issues of the problem statement. Lou Berger strongly supported this, noting that email exchanges were insufficient for resolving these architectural discrepancies. The Chairs agreed to explore scheduling an interim meeting.

3. Babel for IEEE 802.11 (Wi-Fi)

Donald Eastlake presented Babel for IEEE 802.11 (Wi-Fi), which explores using the Babel routing protocol as a Layer 2 path selection protocol within an IEEE 802.11 mesh.

Technical Details:
- Integration involves using vendor-specific elements containing the IETF Organizationally Unique Identifier (OUI) inside mesh profile beacons to indicate Babel path selection.
- Babel would need modifications to route using MAC addresses instead of IP prefixes. Feedback from Juliusz Chroboczek indicated that these changes are significant enough that the resulting protocol might need to be distinguished as "Babel L2".
Discussion:
- MAC Address Collisions: Fred Templin raised concerns regarding MAC address duplication in a Layer 2 routing scheme. Donald Eastlake noted that while MAC duplication can cause issues, it is a general problem inherent to any L2 mesh (such as existing HWMP deployments). Henning Rogge emphasized that MAC address collisions completely break Layer 2 communication, rendering Layer 3 non-functional regardless of IP uniqueness.
- Charter Constraints: The Chairperson noted that incorporating this work would likely require a charter update and coordination with AD Jim Guichard.

4. MANET Awareness Model

Henning Rogge presented the MANET Awareness Model.

Proposed Model: A high-level, protocol-independent data model (utilizing HTTPS, REST/RESTconf, and JSON) designed to provide a unified network topology view, service discovery registration, and geolocation data (particularly for mobile platforms like drones).
Discussion:
- Data Model Duplication: Rick Taylor and Lou Berger cautioned against designing an entirely new model from scratch. Rick Taylor suggested evaluating whether a high-level management function could simply distill this information from existing, detailed IETF YANG models.
- Protocol Independence: Henning Rogge clarified that existing IETF YANG models are highly protocol-specific (e.g., specific to OLSRv2). The goal of this model is to serve as a generic, protocol-agnostic summary interface for client applications.
- DLEP and Topology Gaps: Lou Berger observed that there is currently no DLEP YANG module in the IETF, which represents a clear gap. He recommended that instead of a fresh start, the author should identify gaps in existing generic topology models and propose augmentations.
- YANG Representation: Lou Berger recommended representing the model using standard YANG tree diagrams. The co-chair noted that the current YANG model already validates successfully using Pyang.

Decisions and Action Items

Action Item: Rick Taylor to review the DLEP RFC 8175 errata on the mailing list.
Action Item: The Chairs will organize an interim meeting to address and resolve the architectural and problem statement concerns in draft-ietf-manet-inet-gap-analysis before the next face-to-face meeting.
Action Item: Participants are requested to review Christopher Dearlove’s newly adopted draft and post feedback to the mailing list.
Action Item: The draft owners for the Babel for IEEE 802.11 (Wi-Fi) work will continue discussion on the mailing list, specifically addressing the MAC address duplication concerns and Juliusz Chroboczek's feedback.
Action Item: Henning Rogge to investigate augmenting existing generic IETF topology models (rather than starting from scratch) and include YANG trees in future versions of the MANET Awareness Model draft.

Automatic IETF Minutes