Minutes IETF110: teas

Meeting Minutes Traffic Engineering Architecture and Signaling (teas) WG
Title Minutes IETF110: teas
State Active
Other versions plain text
Last updated 2021-07-25

Meeting Minutes

# TEAS Agenda For IETF 110
Version: Feb 24, 2021

Tuesday, March 9 2021
17:00-19:00 Session III (UTC+1)
Time Zone Converter:   

Materials:      https://datatracker.ietf.org/meeting/110/session/teas
Note taking:    https://codimd.ietf.org/notes-ietf-110-teas
Audio stream:   http://mp3.conf.meetecho.com/ietf110/teas/1.m3u Jabber:        
http://jabber.ietf.org/logs/teas WG  ICS:       
https://datatracker.ietf.org/meeting/upcoming.ics?filters=teas Session ICS:   

YouTube:    https://www.youtube.com/watch?v=jahmWi0aRRQ

## Slot#   Start Timemin    Duration      Information
## 1       17:00   10min    Title:  Administrivia & WG Status
Presenter:    Chairs

## 2       17:10   0min     Title:  WG Draft updates (discuss on list)
Draft:  Many
Presenter:    Chairs

Vishnu Pavan Beeram: Chairs are debating on whether
draft-ietf-teas-pcecc-use-cases document should be taken to WG LC

Dhruv Dhody: As someone who holds the pen on this, please do give me a headsup
if the plan is to take it to WG LC; There are a few TBD items that need to be
taken care of before starting WG LC.

## 3       17:10 (17:08)  10min    Title:  YANG models for VN/TE Performance
Monitoring Telemetry and Scaling Intent Autonomics / Traffic Engineering (TE)
and Service Mapping Yang Model
Presenter:    Dhruv Dhody

Tarek Saad: There is a model to realize SR policy and a separate model for TE
tunnels. We originally thought that the TE tunnel could be used to realize SR
policy but the SPRING WG decided that the SR policy is not analogous to TE
tunnel. Is that what is triggering the issues here?

Dhruv Dhody: When we started the VN yang model, we decide that we would reuse
the TE tunnel and TE topology models as much as we could. If we have to now
also map to the SR policy, it makes our model a bit ugly. Not sure if that is
the best way forward and it is worth discussing more.

Tarek Saad: Agree that this needs more discussion. We initially did add SR path
specific attributes like color information to the TE tunnel, but we removed all
of that after the SPRING WG decision. This discussion should be resurrected.

Dhruv Dhody: The SR policy model is lacking most of the constraints information
that are already defined in the TE tunnel. The constraints should be as similar
as possible for SR policy. From the point of view of an NBI that is used for
requesting path placement, the path control technology that is being used --
RSVP or SR -- should not matter. Should we do a profile of TE tunnel similar to
the profile for TE topology?

Oscar Gonzalez de Dios: Currently SR policy is only a device mode, not
applicable at the NBI of a controller.

Vishnu Pavan Beeram: The SR policy model should be using some of the TE
groupings that we have developed in this WG. This is an action item for the
authors of the SR policy model. For the ACTN VN YANG model, there are a couple
of things that you need to worry about -- one is the underlay topology and the
other is the underlay path. For the underlay topology it should be possible to
refer to the sr-te-topology draft that is being discussed in our WG; but for
the SR path we do have an issue because the generic TE path modeled in TEAS
does not directly map to the SR path.

Vishnu Pavan Beeram: Please take it to the list, this is something that needs
to be resolved.

## 4      17:20  (17:21) 5min     Title:  Interworking of GMPLS Control and
Centralized Controller System Draft: 
Presenter:    Yi Lin

Lou Berger: I've noticed that you do not cover FRR. You would need to cover
that. Yi Lin: Yes, that can be done.

## 5      17:25  17:?? 20min    Title:  Definition of IETF Network Slices
Presenter:    Reza Rokui

Lou Berger: The document is a WG doc, so any agreement has to be at WG level
not among authors.

Joel Halpern: The slide doesn't list all the discussed options.

Reza Rokui: We will revise to capture all options discussed on the list.

Italo Busi: There is a difference between AP and VNAP in RFC8453. The AP is the
reference point between the customer and the provider and can support one or
more VNAP. The VNAP is the binding between a VN and an AP.

Reza Rokui: The NS-AP has the same meaning.

John Drake: When do the chairs intend to make a call on the terminology?

Lou Berger: The chairs are tracking the discussion on the list and think there
is still some room for discussion.

Eric Gray: Terminology needs to be aligned in order to facilitate smooth
merging of the definitions document and the framework document.

Lou Berger: Chairs have the same view.

## 6      17:45 (1746)   10min    Title:  Realizing Network Slices in IP/MPLS
Networks Draft: 
Presenter:    Tarek Saad / Vishnu Pavan Beeram

Adrian Farrel: The notion of slice policy needs to be clarified further.

Tarek Saad: This is explained in subsequent slides.

Chenhao Ma: The new terms Slice Aggregate and Slice Policy are a little
confusing, is there any difference from the term VTN in VPN+ framework?

Tarek Saad: There are some key differences between a Slice Aggregate and VTN.
The main difference is that Slice Aggregate is NOT a topology.

Zhenbin Li: There are many approaches (some overlapping) for carrying the slice
identifier in packets. There is a need to find some consensus.

Tarek Saad: Agree that there needs to be some consensus with respect to the
common pieces of functionality.

Lou Berger: It may be useful to have some of these generic constructs in the
framework document.

## 7      17:55 (1759)  5min     Title:  IETF Network Slice Controller and its
associated data models Draft: 
Presenter:    Luis M. Contreras

Lou Berger: if you think there is some material in this draft which can be
useful for the framework draft, please consider proposing on the mailing list
some text that can be added to the framework draft.

## 8      18:00   8min  (1805)   Title:  A Yang Data Model for IETF Network
Slice NBI Draft: 
Presenter:    Reza Rokui

Lou Berger: It is clear that we need an NBI draft. It is good to see this

## 9      18:08   8min   (1813)  Title:  YANG Data Model for Slice Policy
Presenter:    Tarek Saad / Vishnu Pavan Beeram

Zhenbin Li: Should the control-plane extensions drafts be discussed before
putting together a data model draft?

Vishnu Pavan Beeram: No, the first step would be to get consensus on the need
for a slice aggregate construct and to model the construct that instantiates
it. All other control plane / data plane extensions can follow suit.

Adrian Farrel (from chat): Terminology is not consistent with
draft-bestbar-teas-ns-packet :-(

Tarek Saad (from chat): Thanks Adrian. Appreciate a pointer to the inconsitency
and we will make sure it is corrected

## 10     18:16   8min (1821)    Title:  5G End-to-end Network Slice Mapping
from the view of Transport Network Draft: 
Presenter:    Xuesong Geng

Uma Chunduri: Lot of relevant work is happening in the DMM WG.

Jie Dong: We are open to discuss which WG this work belongs to. We can take
this discussion offline.

## 11     18:24   (1829) 8min     Title:  Instantiation of IETF Network Slices
in service providers networks Draft: 
Presenter:     Luis M. Contreras

No questions/comments

## 12     18:32 (1838)  8min     Title:  IETF Network Slice use cases
Presenter:     Ruixue Wang for Weiqiang Cheng

Luis Contreras: Suggest to check consistency of use cases presented here with
what are in the draft-contreras-use-case draft mentioned in the previous

## 13     18:39  (1244) 8min     Title:  Scalability Considerations for
Enhanced VPN (VPN+) Draft: 
Presenter:     Jie Dong

No questions/comments

## 14     18:46   8min     Title:  Profiles for Traffic Engineering (TE)
Topology Data Model Draft: 
Presenter:     Italo Busi

Vishnu Pavan: Please capture the questions from the last slide and send it to
the mailing list to get feedback.

Italo: Sure.

## Adjourn 19:00