Session Date/Time: 21 Mar 2024 23:30

nmrg Meeting Minutes

Summary

The Network Management Research Group (NMRG) session covered updates on existing documents, presentations on emerging network management topics, and discussions on collaboration and future directions. Key topics included semantic communication networking, distributed zero-trust architecture, intent-based network management, network digital twins, and optical network telemetry.

Key Discussion Points

Document Updates: Two documents were in last call, one on challenges and opportunities for networking and the other on research challenges in coupling AI and network management. The AI/network management document requires a short final call for comments due to a new version submission.
Semantic Communication Networking: A presentation discussed the application of machine learning models to semantic communication networking, drawing parallels with linguistics and pragmatics. The focus is on improving semantic and effectiveness levels of communication.
Distributed Zero-Trust Architecture: The session included a discussion on implementing zero-trust architecture in fully distributed environments, using federated learning and blockchain.
Intelligent TCP Protocols: Rethinking TCP protocols by plugging in machine learning for intelligent networking was discussed. The potential of experience-driven multiparty TCP was highlighted.
Joint Exposure of Network and Compute Information: A presentation explored the merging of communication and compute, particularly at the edge, and the need for a structured way to expose compute and communication information from the network. This included use cases such as hybrid AI and distributed XR computation. Alignment with other groups such as CATS and ALTO was suggested.
Intent-Based Network Management (IBN): A draft proposing an IBN architecture with specific, measurable, achievable, rewarding, and time-bound characteristics was presented. A demo of an intent-driven service function chain was shown.
Interconnection Intent: An update on a draft concerning interconnection intent in multi-domain environments was given, focusing on automation and resource exchange beyond IP.
Software Defined Vehicle (SDV) Networking: A presentation outlined an intent-based management framework for software-defined vehicles and intelligent transportation systems.
Intent Translation Engine: A presentation described an intent translation engine for translating high-level user intents into network API and cloud API configurations.
IBN Use Case Integration: Discussions focused on how to collaborate on and integrate different IBN use cases, potentially merging them into a consolidated document with a common outline. It was noted there are different types of outcome targetted by each use case.
Network Digital Twin (NDT): A presentation provided an update on the network digital twin draft, including a discussion of key characteristics and a sample use case related to 5G network slicing. Concerns about control interfaces and the scope of the draft were addressed.
Optical Network Telemetry: A presentation explored telemetry for optical networks, focusing on physically oriented data and use cases related to optical network digital twins and fault management. The need for continuous-time, real-number measurements and preprocessing techniques was emphasized.

Decisions and Action Items

AI/Network Management Document: A short, two-week final call for comments will be initiated for the AI/network management document.
IBN Use Cases: Authors of different IBN use cases will further discuss integrating their documents into a consolidated draft.
NDT Architecture: A design team will be established to focus on the network digital twin architecture, with regular online meetings to address key issues. Splitting the draft into concept/definition and architecture sections will be considered.

Next Steps

Chairs will track action items to ensure timely completion.
The IBN use case authors will continue discussions on integration.
The NDT design team will begin focused activity on the architecture.