Minutes IETF99: nmrg: Thu 18:10
minutes-99-nmrg-201707201810-00

44st Network Management Research Group (NMRG) meeting
IETF-99
July 20th 2017 - Prague, Czech Republic


18:10 - 19:10 : Thursday Afternoon session III - Karlin III
-----------------------------------------------------------

18:10 - 18:15 : Agenda Bashing
                (Lisandro Zambenedetti Granville)

-----------------------------------------------------------

18:15 - 18:25 : Intelligent management using collaborative reinforcement
                multi-agent system
                (Minsuk Kim)

Jeferson Nobre - Are you referring in the presentation to autonomous
                 driving system or an autonomous automotive networking?
Minuk Kim - We employed for the proposal as a whole
Jeferson Nobre - What are the used models? Yang?
Minuk Kim - We use our own data models.
Lisandro Granville - What is the difference between your work and the
                 work in the 2000's? You should review those work to
                 avoiding occasionally proposition a solutions that has
                 been represented before.

-----------------------------------------------------------

18:25 - 18:35 : A Reference Model for Managing Large-Scale Software-Defined
                Network (SDN) Infrastructures
		(Klaus Wehmuth)
                https://datatracker.ietf.org/doc/draft-wehmuth-nmrg-sdn-model/
                draft-wehmuth-nmrg-sdn-model-00

Lisandro Granville - Which is the version of the draft?
Klaus Wehmuth - First version.
Lisandro Granville - Did you receive feedback?
Klaus Wehmuth - No.
Diego Lopez - What are possible results for the proposal? 
Klaus Wehmuth - Calculate distances, Representation the net, etc.
Diego Lopez - It would be important to have tools that use the proposal,
              for example, some people use eclipse to validate the networking
              models.
Klaus Wehmuth - We can do that, only if one needs special results. You just
              have to add this to the proposed algorithm, for example for
              projection of the vertices, or another flow algorithm.
Diego Lopez - Are there dynamic matrices?
Klaus Wehmuth - Right.
Diego Lopez - Are you using any graph language?
Klaus Wehmuth - No, but we can work to bring this, I think it is possible.
Diego Lopez - Do you have any test or preference for the proposal?
Klaus Wehmuth - Our approach is used on different kinds of networks, for
              example transportation, not in the computer network.
Diego Lopez - It is good if the draft have links for the employed tools.

-----------------------------------------------------------

18:35 - 18:45 : Framework for Autonomic Management in Software Defined Networks
                (Angela Rodríguez-Vivas)

Jeferson Nobre - It is important to connect the proposed work with the IETF
               body of work in the area.
Angela Rodríguez-Vivas - Thanks for the suggestion.

-----------------------------------------------------------

18:45 - 18:10 : Discussion

Alex Clemm - There is a lot of talk about intent, but what it is really?
             There is , differences between previous models, some seems
             closely related. It is helpful to develop a position, which
             can avoid different terminology. Policies are not just ECA,
             intent can be viewed as a high level policy. , It is important
             to understand how these policies will be achieved, maybe
             policies can be more like programmed thing. I think it is good
             to have position or guideline about intent in the NMRG.
Dan Romanascu - There are other works in IETF about intent.
Lisandro Granville - Yes, and the approach is to recognize those works but
             put them in perspective with related topics, e.g., PBNM.
Dan Romanascu - A taxonomy or something would be helpful.
Jeferson Nobre - ANIMA "sidelined" this in the charter.
Bert Wijnen - Every intent is a policy, but not all policies are intents.
Alex Clemm - There was much discussion in the 2000s about policies,
             informative models, etc.
Dan Romanascu - There is an RFC about this.
Lisandro Granville - The suggestion is to proceed the discussion in the
             mailing list about intents, specially in the implied research
             challenges or other material that could be brought to the NMRG.