Hi,

Comments inline.

--
David Harrington
[email protected]
+1-603-828-1401


On 9/14/12 9:54 AM, "t.petch" <[email protected]> wrote:


>
>With any I-D/RFC, you either get it adopted by a Working Group or you do
>it as an individual submission.  In terms of your work, both are about
>equal.  In terms of the work of others, the latter tends to burden some
>people more than the former so there is a push to use a WG.

I see things differently.
I think the individual draft requires less work by others.
An AD can sponsor the draft, and ask the IESG to approve it.

A WG is used to get a wider review of the proposal.

>There is no
>WG currently in the IETF that focuses on MIBs, this, OPSAWG, is about
>the nearest.  The key players are the WG chairs and an Area Director;
>the web site tells you who they are and you can see on the WG list
>archive what they have said lately about what and form a view as to how
>supportive they are.  Well, ask them directly if that is your style.
>Look too at the WG charter
>
>"The OPSAWG will undertake only work items that are proved to have at
>least a reasonable level of interest from the operators and users
>community and have a committed number of editors and reviewers."

I have not seen a large demand for this proposed object from operators.
I think asking the WG for review will be important to
1) establish operator demand,
2) establish implementation feasibility of providing this information via
SNMP,
3) help to ensure that a clear, unambiguous definition of "available"

>
>Um.
>
>On the route of individual submission, it as an AD and a WG chair (or
>perhaps a former WG chair) who have a lot more work to do to push an I-D
>through to a standard; you would need their active support.  Hence the
>push to use a WG - and if the WG and/or its chairs are reluctant to take
>it on, then the work is unlikely to happen.

IIRC, a WG chair is NOT needed to push an ID through to standard.
An AD can sponsor it alone, although most like to see a separate document
shepherd (who might be a chair or ex-chair).
The AD just needs to convince the IESG to approve it as a standard.
And the IESG might want some proof of operator demand, implementation
feasibility, and community consensus on the definition.

>
>As you probably gather, SNMP and MIB modules are regarded somewhat as a
>legacy technology, Netconf/Netmod being the current incarnation.  So if
>you were to propose the conversion of the hr MIB module to netmod, then
>I would expect you to see more enthusiasm; but that is, of course, a
>quantum leap.


Since the object being proposed reports operational state, and is a
read-only object that is not configurable, and detection of significant
state change may be important for applications, periodic polling of a MIB
object using SNMP seems the appropriate choice of tool; I don't see this
as configuration information appropriate for the netconf/YANG toolset
(except as part of an overall HR configuration data model - but that's a
whole different proposal). Therefore I would not expect more enthusiasm
for a netconf/YANG solution.

>
>Tom Petch
>
>> Thanks again,
>> Sheppy
>>
>
>
>_______________________________________________
>OPSAWG mailing list
>[email protected]
>https://www.ietf.org/mailman/listinfo/opsawg


_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to