Re: discontinuity timers for Counters

[Juergen Schoenwaelder <j.schoenwaelder@iu-bremen.de>]

On Tue, Jul 27, 2004 at 02:10:30PM -0700, Randy Presuhn wrote:

> This seems contrary to the whole motivation for using a zero-based
> counter.  I'd think that if they *really* need a zero-based counter,
> that the appropriate semantic is for its value to represent the number
> of events since the beginning of the epoch defined by its discontinuity
> indicator.  If that is so, then the appropriate discontinuity behaviour
> would indeed be a reset to zero.

I tend to agree. We surely lack guidelines when these zero based
counters are appropriate and how they behave when there is a
discontinuity event.

> What bothers me more is that it specifies TWO discontinuity indicators.
> I'd read that as an implicit requirement for management applications
> to retrieve BOTH with every query, in order to determine whether either
> gives evidence of a discontinuity.

But this is what I read out of the IF-MIB RFC section on
discontinuities. Would be interesting to figure out how many
implementations actually do this.

/js

-- 
Juergen Schoenwaelder		    International University Bremen
<http://www.eecs.iu-bremen.de/>	    P.O. Box 750 561, 28725 Bremen, Germany