RE:redundent objects in phivCountersCountEntry Tue, 15 December 1992 14:20 UTC

Received: from by IETF.CNRI.Reston.VA.US id aa01828; 15 Dec 92 9:20 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa01823; 15 Dec 92 9:20 EST
Received: from by CNRI.Reston.VA.US id aa06352; 15 Dec 92 9:22 EST
Received: by; id AA02638; Tue, 15 Dec 92 06:22:03 -0800
Received: by; id AA08785; Tue, 15 Dec 92 06:14:02 -0800
Received: by; id AA08781; Tue, 15 Dec 92 06:14:01 -0800
Received: by; id AA02019; Tue, 15 Dec 92 06:13:59 -0800
Received: by (5.57/ULTRIX-fma-071891); id AA03810; Tue, 15 Dec 92 09:16:59 -0500
Message-Id: <>
Subject: RE:redundent objects in phivCountersCountEntry
Date: Tue, 15 Dec 92 09:16:57 -0500
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
X-Mts: smtp


Sorry for the delay in responding to questions about the duplicate


I remember that we discussed the need for these counters during the early
stages of the MIB development.  In fact, John Shriver sent out a note
explaining the distinction between counting information for DECNet
LINKs and LINEs.

Clearly what has happened is that during the editing stages we messed up
and it was not caught.  While we did preserve the Blocks and Blks
entries to have separate objects for LINK and LINE Data Blocks sent and
received, we did not catch the fact that the OBJECT DESCRIPTIONS were
the same. 

Given that the group has the following conformance language:

       -- Counters Group

       -- The implementation of the Counters Group is mandatory for
       -- systems which only support DECnet style locking counters.

I need to ask, has anybody implemented these objects?  Based on the
answer to this question, I see the following options:

       1.  If nobody has implemented the group, and there is a clear
       consensus that it has proved to be not useful, we can eliminate
       the entire group and this problem along with it.

       2.  If people do want to keep the group, I would be interested to
       learn if anybody has implemented it (or are planning to implement
       it) and if so, what was done with these objects.  One possibility
       is to just keep one pair and only count, for example, link data.

       3.  It is also possible to obsolete two objects and define two
       new ones with different descriptions which I am sure we could
       agree to if necessary.  


	Jon Saperia, Digital Equipment Corporation