"Dan" == Romascanu, Dan (Dan) <dromasca(_at_)avaya(_dot_)com> writes:
Dan> Sam, Thank you for your review and opinions.
Dan> I would like to remind you and let many people that are not
Dan> aware about the history of the document know one fact that
Dan> may be important. This document is an outcome of the
Dan> discussions hold at the IESG retreat in May 2006. I was then
Dan> the 'fresh' AD bringing this issue to the IESG table, we
Dan> discussed approaches on dealing with management in the IETF
Dan> and the need for a different approach of looking at
Dan> management than the 'write a MIB' which was the rule in the
Dan> IETF WGs until then. I took the action item to 'write a
Dan> draft' on this issue - which then developped in this piece of
Dan> work chartered in the OPSAWG.
I certainly appreciate the work that has gone into this draft. I'm
not sure why the origins here are important. If you're saying that it
should have special status because the original discussion happened at
the IESG level, I disagree. If you're saying that the content has
broad consensus because it started at the IESG level, I disagree. If
you're saying that it's important work with a long history, I agree.
Dan> I believe that the document recognizes the variance in
Dan> approaches for the different areas, protocols, and working
Dan> groups in the IETF
I strongly disagree that it succeeds in this goal.
I agree it tries.
As an example, section 3.1 says that the primary goal when considering
management should be interoperability.
That's a broad statement that does not have IETF consensus and is inappropriate
for a BCP.
Dan> and for this reason rightly avoids making
Dan> a prescription or imposing a fixed solution or format in
Dan> dealing with operational considerations and manageability
Dan> aspects of the IETF protocols. I think that it does make
Dan> however the point that operational deployment and
Dan> manageability aspects need to be taken into considerations
Dan> for any new IETF work. The awareness of these issue should
Dan> exist in any work the IETF engages with, after all we develop
Dan> technologies and protocols to be deployed and operated in the
Dan> real life Internet, not abstract mathematical models. It is
Dan> fine if a WG decides that its protocol needs not
Dan> interoperable management or no standardized data model, but
Dan> this should be the result of discussions and decisions, not
Dan> of mission.
It's not at all clear to me from this document that would be fine.
That's one of my most serious problems with the document.
Ietf mailing list