wiki:WikiStart

Network Management Research Group (NMRG)


Introduction

The Network Management Research Group (NMRG) of the Internet Research Task Force (IRTF) was approved by the Internet Architecture Board (IAB) on Sunday, 14 March 1999. If you have any questions, comments, or wishes, feel free to contact the research group by sending email to the nmrg@… mailing list. The mailing list is being archived at http://irtf.org/mailman/listinfo/nmrg.

Charter

The Network Management Research Group (NMRG) provides a forum for researchers to explore new technologies for the management of the Internet. In particular, the NMRG will work on solutions for problems that are not yet considered well understood enough for engineering work within the IETF.

The focus of the NMRG will be on management services that interface with the current Internet management framework. This includes communication services between management systems, which may belong to different management domains, as well as customer-oriented management services. The NMRG is expected to identify and document requirements, to survey possible approaches, to consider new architectural frameworks, to provide specifications for proposed solutions, and to prove concepts with prototype implementations that can be tested in large-scale real-world environments.

The IETF Operations and Management Area Directors are members of the NMRG mailing list and invited to NMRG meetings in order to ensure free flow of information in both directions, and to avoid duplication of work with the various IETF working groups.

The group will report its progress through a publicly accessible web site and presentations at IETF meetings. Specifications developed by the NMRG will be submitted for publication as Experimental or Informational RFCs.

Membership

Membership in the NMRG is open to all interested parties.

Meetings

Regular working meetings are held about three to five times per year at locations convenient to the majority of the participants. Working meetings vary from hours-long working sessions (typically when held as part of IETF meetings) to days-long meetings when co-located with conferences or events related to network management. Regular virtual meetings are also organized on a monthly or per-need basis.

Research Activities (2017-2022)

The constant evolution of networking technologies, in scale, versatility, and heterogeneity, generates operational complexity and demands novel disruptive management solutions to address it. The NMRG will prioritize investigation of three related topics: 1) self-driving/-managing networks, 2) intent-based networking and 3) artificial intelligence in network management. Note: beyond these three topics, the NMRG remains open to presentation of other topics of interest.

While the ultimate goal of self-driving/-managing networks is fully autonomous network operations, there will be intermediate levels where the human users remain “in the loop” and are progressively assisted and replaced by more and more intelligent mechanisms. Interfaces between humans and a self-driving system are important and required to allow bidirectional communications. On one hand, the user must be able to express guidance and its needs without having to handle the full complexity of the underlying infrastructures. On the other hand, users must understand the decisions which were taken and the reasons why, be informed about the future actions the system will initiate and also be provided with recommendations.

In this area, Intent-Based Networking (IBN) provides high-level, user-friendly abstractions to describe business and operational goals, and alleviates the need for the user to know and derive the technical details on how to achieve those goals. IBN is an essential component of self-driving networks but requires the introduction of intelligent mechanisms to properly process intents with as little human involvement as possible.

Certainly, some of those intelligent mechanisms can rely on advances in (but should not be limited to) Artificial Intelligence (AI). While different forms of AI have been used for decades in network management, the combined progress in amount of data, computing power, AI algorithms and flexible capabilities of networks in recent years makes highly relevant to re-examine in depth the coupling between AI and network management.

Work plan

To investigate these topics, the initial set of work items comprises:

For Intent-Based Networking (IBN):

  1. Document the problem statement, design goals and challenges.

Goal: describe the problem and solution spaces; identify the limits of current technologies and methods and derive the associated research challenges.

  1. Document fundamental concepts, background, and terminology.

Goal: provide clarity and achieve a common understanding of the various concepts, definitions and terms of what constitutes an IBN system.

  1. Develop a taxonomy and document suitable means to express intents.

Goal: categorize the different forms of intents and define what constitutes a “well-formed” intent; describe how an intent can be expressed and what can be expressed using an intent with means such as information models, grammars, and languages.

  1. Design and specify a common architectural framework comprising requirements, functions and techniques to realize an archetypal IBN system; describe the lifecycle and theory of operations.

Goal: determine the elementary functional blocks of an IBN system, their interactions, inputs and outputs; propose different techniques applicable for the different functions.

  1. Define appropriate validation scenarios and use cases describing concrete examples of intent expressions and functions.

Goal: assess the quality and completeness of specifications and evaluate intent-based systems functionalities in experimental settings.

  1. Develop implementations and proof of concepts.

Goal: demonstrate the feasibility of the proposed framework and its functions; detect potential design flaws, and provide a basis for interoperability evaluations.

  1. Study the integrability and interoperability aspects of the proposed IBN architectural framework.

Goal: enable the large adoption and applicability of IBN with existing and emerging technologies, and provide guidance on deployment considerations.

For Artificial Intelligence in Network Management (AI-NM):

  1. Investigate, organize and document the major research challenges in AI for Network Management.

Goal: provide a reference document which defines the different forms and usages of AI in network management and articulates the different goals, challenges, requirements and research directions.

  1. Organize and animate a series of practical Network Management AI challenges/competitions.

Goal: promote experimental research, practical knowledge and validation of AI techniques to solve network management problems and foster exchanges and cross-participation of both AI and Network Management specialists.

  1. Support discussion and collaboration on techniques, (meta-)data, experimentations and best practices for the use and integration of AI with networking management approaches.

Goal: offer a forum for the Network Management AI community to report on advances, developments and key results and introduce its efforts to the IETF. Note: Applicability of AI techniques for IBN functionalities and mechanisms is an example of potential joint activity between the Network Management AI and IBN realms.

For Self-Driving/-Managing Networks (SD/MN):

  1. Support discussion to develop a common understanding of the problem-solution space on new architectural frameworks, articulate related requirements, survey and propose possible novel approaches.

Goal: offer a venue for the Network Management community to debate on current Internet management frameworks and new proposals, and how to adapt and anticipate on needs, technologies and ecosystem evolution.

  1. Investigate and document reference models and de-facto best practices.

Goal: describe how various realms and components, such as intent-based functionality, automation and zero-touch capabilities, or else algorithmic approaches (AI or non-AI based), compose together to form modern, comprehensive and coherent management solutions.


Milestones

Note: progress on AI-NM work item 3 (research forum) and SD/MN work items 1-2 will be reported on a regular basis, or at specific event (e.g. meetings co-located/-organized with conferences). The outcome of these work items will not necessarily generate stand-alone documents, nor specific milestones, and thus are not listed explicitly in the list of milestones.

November 2019

  • Document(s) for IBN work items 1-3 submitted to RG adoption.
  • Preliminary results on IBN work items 5-6 communicated.

March 2020

  • Document(s) on AI-NM work item 1 submitted to RG adoption.
  • Framework for AI-NM work item 2 published.

July 2020

  • Document(s) for IBN work item 4 submitted to RG adoption.
  • Intermediate results on IBN work items 5-6 communicated.

November 2020

  • Document(s) for IBN work items 1-3 submitted to IRSG review.
  • Document(s) on IBN work items 5-6 submitted to RG adoption.
  • First set of results and report on AI-NM work item 2 published (i.e. Challenge #1).

March 2021

  • Document(s) for IBN work item 7 submitted to RG adoption.
  • Document(s) on AI-NM work item 1 submitted to IRSG review.

July 2021

  • Document(s) for IBN work item 4 submitted to IRSG review.
  • Second set of results and report on AI-NM work item 2 published (i.e. Challenge #2).

November 2021

  • Document(s) on IBN work items 5-6 submitted to IRSG review.
  • Document(s) for IBN work item 7 submitted to IRSG review.



NMRG Co-Chairs

  • Jérôme François
  • Laurent Ciavaglia



Resources


[These NMRG wiki pages have been ported from the original previous NMRG website (http://www.ibr.cs.tu-bs.de/projects/nmrg/) by Arthur Selle Jacobs (asjacobs@…), Ramon Costa Silva (ramoncsilva@…) and Felipe Barbosa Tormes (fbtormes@…)]

Last modified 3 months ago Last modified on 23/03/20 10:10:21