[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: RowStatus versus RowState


"Ayers, Mike" wrote:
> > From: Sharon Chisholm [mailto:schishol@nortelnetworks.com]
> > The rowops draft says that the intention is to replace RowStatus with
> > RowState, but that RowStatus is not being deprecated.  I think this
> > is a bit ambiguous.
>         It's worse than that.  The current intention appears to be to have
> RowState do only part of the job.  Specifically, there has been, both in the
> rowops draft and in the proposed row ops requirements, text that claims that
> "state and fate" are separable entities.  Really, all that has been done is
> that half of the RowStatus values have been converted to "rowops" and the
> timeout on notReady has been removed - presto chango!

Why do you think that RowState+RowOps does not
achieve the "fate vs state" claim?  Seems to me
this claim is basically realized.

>         What is probably more iomportant at this point is to make progress
> on the requirements so that we can determine what problem, if any, we are
> trying to solve.  I still fail to see any point to rowops.
> /|/|ike

As you noted, a list of POTENTIAL and sometimes CONFLICTING rowops
requirements was posted a few weeks ago along with a request
for comments.  So far very few, if any, comments have been received.

The current rowops draft reflects a subset of those requirements.
In sum, this draft attempts to make do with current PDUs and structures,
provide reasonable (if not optimal) packing efficiency, separate fate
from state :-), not impact current VACM/AgentX configurations (though
AgentX itself MAY have to evolve), allow a reasonable and incremental
transition strategy, provide opportunities to simplify NMS appl and SNMP
agent instrumentation coding efforts (though not necessarily master
or other core code modules), provide some new features (bulk head/tail,
extended errors, three-phase SETs), etc..

I agree with you that more discussion is needed regarding requirements.
This draft simply serves as a backdrop from which we can compare
ideas.  For example, Steve and Sandra have proposed an interesting new
of aggregating rowOps in existing PDUs.  At this point we need more
list discussion/consensus before we can make wise choices and produce
a more ideal draft.

Thanks, Lauren