Re: PerfHist-TC-MIB

["C. M. Heard/VVNET, Inc." <[e-mail]>]

On Wed, 3 Jun 1998, Jeff Johnson wrote:
[ ... ]
> xyzValidIntervals + xyzInvalidIntervals may not equal the number of
> intervals, mainly because xyzValidIntervals is a misnomer.  It is really
> the index of the oldest valid interval.  Consider a box that has been
> running for a long time, but where many intervals are invalid for whatever
> reason.  If interval 1 and 96 is valid and intervals 2-95 are invalid,
> xyzValidCount will be *96* and xyzInvalidCount will be 94.

After carefully re-reading the description clauses for the various
xyzValidIntervals I see that this is the case.   However, the original
question was:

> 2) The xyzInvalidIntervals DESCRIPTION states simply "The number of
> intervals for which no valid data is available."  When an agent which
> supports 96 intervals is initially powered up, should the invalid interval
> counter be 0 or 96?  In other words, are there initially 96 intervals which
> are all invalid, or are there zero intervals?

The description of sonetMediumInvalidIntervals in the most recent SONET-MIB
(see <draft-ietf-atommib-sonetng-02.txt>) is

 "The number of intervals for which no valid data is available.
  This includes intervals of which valid data has been lost,
  e.g., due to restart of the agent."

It seems that the correct answer to this question is that there are zero
intervals, but based on this description I had concluded that the amswer
should be that there are 96 intervals, all invalid owing to the restart
of the agent.  So I think that a clarification is needed.  I would agree
with David Fowler's statement that

> All we really need to do is clarify that the number of invalid intervals must
> be less than the number of valid intervals and that it is intended for the
> proxy situation.

Mike
--
C. M. Heard/VVNET, Inc.
[e-mail]