RE: guidelines section 4.6.1.2 (one hour rule)

["C. M. Heard" <heard@pobox.com>]

On Wed, 5 Feb 2003, C. M. Heard wrote:
> On 28 Jan 2003, Juergen Schoenwaelder wrote (on
> comp.protocols.snmp):
> > > The RFC 2578 mentions that one should use Counter64 only if Counter32
> > > would have wrapped in an hour. Why is this so?
> > 
> > There are recently been some debate about this in the IETF and I guess
> > this rule does apply in this form anymore. There will be a MIB reviewer
> > guidelines document soon which will address among many other things this
> > issue. In fact, there is already a published IETF MIB in RFC 3289 which
> > solely uses Counter64 and I guess we will see more of them in the
> > future.
> 
> I missed that one.  It was not on my radar screen :)

On Wed, 5 Feb 2003, Wijnen, Bert (Bert) replied:
> Mmm... I thought i had mentioned several times allready
> that it is OK with me to specify just Counter64 if
> that is what a WG thinks makes sense. Certainly if 
> some of the counters in their MIB module are Counter64
> anyway, then I have no problem if they make them all
> counter64. 
> 
> If the roll-over of a counter32 is LT one hour, I 
> would tell them they MUST use Counter64.

OK, here is some proposed text.  If accepted, it would be inserted
just before the next-to-last paragraph of section 4.6.1.2:

   RFC 2578 Section 7.1.10 places a requirement on "standard" MIB
   modules that the Counter64 type may be used only if the information
   being modeled would wrap in less than one hour if the Counter32 type
   was used instead.  Now that SNMPv3 is an Internet Standard and SNMPv1
   is Historic (see [RFC3410]) there is no reason to continue enforcing
   this restriction.  Henceforth "standard" MIB modules MAY use the
   Counter64 type when it makes sense to do so, and MUST use Counter64
   if the information being modeled would wrap in less than one hour if
   the Counter32 type was used instead.

If I don't hear any objections, I'll put this in the next version of
the document.

//cmh