Gentle rtgwg and ops-area participants,
I wanted to call attention to a couple of drafts about early activity on a YANG
profile for data center switches that will be discussed in RTGWG (first
session, Wed , 9a start) and OPS-AREA (Thursday, 1p start):
https://datatracker.ietf.org/doc/html/draft-wbl-rtgwg-yang-ci-profile-bkgd
https://datatracker.ietf.org/doc/html/draft-wbl-rtgwg-baseline-switch-model
For those who don't want to wade through the drafts, the RTGWG chairs have
already posted the slides:
https://www.ietf.org/proceedings/98/slides/slides-98-rtgwg-yang-device-profile-for-redfish-network-management-draft-wbl-rtgwg-baseline-switch-model-draft-wbl-rtgwg-yang-ci-profile-bkgd-00.pdf
[watch out for URL split across lines ...]
The short summary is that for data center converged infrastructure [CI],
networking should be managed as part of the overall infrastructure - e.g.,
single "pane of glass" for an 84U rack of which 4U may be networking equipment.
Another standards body (DMTF - Distributed Management Task Force) is working
on a standard for CI management, Redfish, and is reusing YANG models for
network management via a translation approach (hallelujah - not-invented-here
doesn't win every time!) - among the benefits of reuse this should be common
views of underling state via YANG-based CI management and network management
tools.
This reuse is working "running code" today, as described in the slides and the
ci-profile-background draft. The next step is to define the minimum
implementation requirements for what data center switches to be managed as part
of CI - i.e., when a YANG-based management tool looks at such a switch, what
does it expect to see as the minimum set of supported YANG modules, i.e., a
device profile. This avoids "defensive programming" in management tools to
deal with different switch implementations supporting different YANG modules -
that inevitably leads to switch-specific code in management tools. The
baseline-switch-model draft is a rather drafty first start on the minimum set
of supported YANG modules for data center switches (NB: the set will differ for
other classes of equipment/devices) - we're looking to do this work in the IETF
so that the resulting device profile is broadly useful well beyond Redfish, and
we (draft co-authors) are looking for all the help, advice, etc. that people
care to offer. In particular, we're not attached to the format of that draft
- it was our first attempt to capture material.
The longer, more detailed discussion of this activity will be in RTGWG - the
shorter OPS-AREA session is intended to be more of an overview of the activity
and why it's useful in the hope of engendering interest. We'll be happy to
take questions on the lists, in the meetings or directly.
Thanks, --David (on behalf of Joe White, David Black and John Leung, the wbl
draft co-authors)
--------------------------------------------------------
David L. Black, Distinguished Engineer
Dell EMC, 176 South St., Hopkinton, MA 01748
+1 (508) 293-7953 Cell: +1 (978) 394-7754
[email protected] <=== NEW ===
--------------------------------------------------------
_______________________________________________
rtgwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/rtgwg
