Re: Chassis MIB

Bob Stewart <rlstewart@eng.xyplex.com> Thu, 03 September 1992 18:09 UTC

Return-Path: <owner-chassismib>
Received: by CS.UTK.EDU (5.61++/2.8s-UTK) id AA03596; Thu, 3 Sep 92 14:09:48 -0400
Received: from xap.xyplex.com by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA03586; Thu, 3 Sep 92 14:09:42 -0400
Received: by xap.xyplex.com id <AA09298@xap.xyplex.com>; Thu, 3 Sep 92 14:08:55 -0500
Date: Thu, 03 Sep 1992 14:08:55 -0500
Message-Id: <9209031908.AA09298@xap.xyplex.com>
From: Bob Stewart <rlstewart@eng.xyplex.com>
To: chassismib@cs.utk.edu
In-Reply-To: Niels Ole Brunsgaard's message of Thu, 3 Sep 92 10:01:58 +0100 <9209030902.AA17840@sony.dowtyns.dk>
Subject: Re: Chassis MIB


>In my model of the Chassis MIB, every agent in the chassis provides a
>view of the Chassis MIB.

This should be allowed, expected, and not required.

>How about defining chasUpTime as the max value of all sysUpTime's in the
>chassis.

Shudder.  This has significant implications for the level of communication
among devices in the chassis.  Although I believe we can assume some, we
shouldn't assume full access to each other's full MIBs.  My model is of
loosely connected systems with some back-door communication channel providing
minimal information about what else is there.  A system that supplies the
Chassis MIB uses that channel to collect what it needs.  Anything that begins
to approach full MIB access should be handled by normal manager to agent
access.  

>I like Keith's scheme with indexType and indexValue. 

I like expensive sports cars and fancy vacations.  That doesn't mean I can
have them.  No one is addressing whether implementations can reasonably supply
such information.

	Bob