Network Management Operations (nmop) WG Agenda - Interim on Cooperation with BBF

• When: 2025-02-26 09:00-10:30 UTC
• Co-Chairs: Benoît Claise & Mohamed Boucadair & Reshad Rahman

Detailed Agenda

1. Introduction (Chairs)

The meeting started with some delay to let colleagues from BBF join.
Benoît went then through the Chairs Slides.

It was reported that some BBF colleagues encountered issues to connect
because of the need to have a datatracker account.

2. BBF LS

https://datatracker.ietf.org/liaison/1969/
https://datatracker.ietf.org/liaison/1975/

Xueyen went through the LS reply from the BBF. The LS also enclosed
the latest version of the WT-508 specification.

Med asked about the target publication date for WT-508 spec. Bruno
Cornaglia indicatated that the document can't be finalized before mid
2026. Benoît commented that date is aligned with the target date for the
message broker (currently scheduled to be submitted to the IESG by Sept
2025).

Xueyan invited the IETF participatnts to attend the forthcoming BBF
Spring meeting that wil be organized in the week right before IETF#122.
Benoît asked about the conditions to attend that meeting. Bruno replies
that the participation can be as a guest (with strong participation
limitations) or via membership. Bruno offered to share the link for
interested participants.

3. draft-ietf-nmop-yang-message-broker-integration Overview

Thomas presented his slides.

Mauro Tilloca raised a question about clarifying the use of YANG
procotol vs data model. Thomas clarified that data is extracted one time
and then consumed multiple times.

Fransisco de Carvalho asked what are the problems that can't be solved
by gRPC/gNMI. Reshad Rahman (via the chat) replied that YANG-Push is an
extension to NETCONF (and RESTCONF) to be able to subscribe (periodic
and on-change) to data.

Fransisco reiteratd his question about what issues YANG Push solves, but
not gNMI. Thomas answered that the various gNMI implementations
available out there are not uniform let alone the lack of key features
for deployment (e.g., capability exchange). Slide#15 also identifies a
list of items that are unique to YANG-Push. That said, there are also
enhancements to YANG-Push (extensibility, in particular).

Reshad asked a question about the client side. Thomas replied that
current management systems are used for that, but a python client is
used for prototype purposes.

Mauro asked about operator backing and also what is the difference
between IPFIX vs. YANG Push. Thomas answered that comparaison is
included in the comparaison slide. Thomas also reminded the list of
operators/vendors supporting this work (see the slides for the list).

Benoît (WG Chair) clarified that typically, in IETF meetings, we try to
respect the agenda slot times. The WG chairs usually enforce this. The
slow start to the datatracker Id did not help. We should have foreseen
this and warned our BBF colleagues. Sorry about that. The BBF/NMOP
cooperation is obviously the end goal of this meeting. However, the
couple of clarifications discussed just now are fundamental for good
cooperation. So we propose to have the right foundation first, to take
the required amount of time for clarifications and therefore not enforce
the slot durations. We will see how we do in term of time towards the
end of the meeting. Either extend the call, continue on the mailing
list, or schedule another interim meeting. No objection was raised
against this proposal.

4. Introduction to WT-508: Scope & Goals

Jian Zhu presented the slides.

5. BBF/NMOP Cooperation, including Some candidate action items for discussion

Thomas went through the slides.

Bruno raised a comment about the scope (push/pull) as the BBF is working
on both. Mauro indicated that the BBF focuses first on the components
before dealing with protocols choice. Ilias Gravas stated both IETF/BBF
arch have a common goal to share and consume data. He also mentioned
that even if in the BBF side they are trying to "abstract" from the
protocols, he supports standardized protocols.

Tom asked whether the broker integration was explored with the use of
IPFIX. Thomas replied the architecture is done to be extensible and thus
be feed with IPFIX.

Fransisco commented that the focus on the BBF side is end-to-end.

6. Wrap-up (Chairs)

Follow-up from the Chairs:

• Consider elaborating a readily-available argumentary about why
  YANG-Push is needed vs. gRPC/gNMI.
• Agree on the scope and expectations of the architecture and level of
  protocol work in both IETF/BBF.

The draft minutes will be shared with the BBF colleagues for
review/comment before publishing the final version.