Re: Integral management agents in a chassis.

Bob Stewart <> Wed, 30 December 1992 20:09 UTC

Return-Path: <owner-chassismib@CS.UTK.EDU>
Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA20621; Wed, 30 Dec 92 15:09:33 -0500
X-Resent-To: chassismib@CS.UTK.EDU ; Wed, 30 Dec 1992 20:09:32 GMT
Errors-To: owner-chassismib@CS.UTK.EDU
Received: from by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA20613; Wed, 30 Dec 92 15:09:13 -0500
Received: by id <>; Wed, 30 Dec 92 18:08:48 -0500
Date: Wed, 30 Dec 92 18:08:48 -0500
Message-Id: <>
From: Bob Stewart <>
In-Reply-To:'s message of Wed, 30 Dec 92 10:11:31 EST <>
Subject: Re: Integral management agents in a chassis.

>How would a chassis with a built-in management agent be represented in
>the chasSlotTable?  It has no slot number.

That's a very good question.  This sort of problem is a pain in various MIBs.
We want nice, simple indexes AND we want them to correspond to physical
reality AND physical reality varies considerably at the implementer's whim.

In other MIBs, this has been handled with a DisplayString that contains a
proper text label for the the slot/port/thingy, and any correspondence with
the index integer is fortuitous.

This problem may be helped if we can figure out what to do about the model
discussion that overwhelmed us in Washington.  If we had types of "slots" and
the type spaces were numbered independently, we could have general-physical
slots 1-8 and virtual-internal-manager-slot 0.  There's something to be said
for the "slot" table being indexed by type and index.  Index could then be
clean, simple, and correspond to silk screening.  Unfortunately type would
have to be an OID, but at least it could be relatively meaningful.

In other words, to answer your question, "Duh.  I dunno."