uptimes and timestamps

bierman@davidsys.com Wed, 12 August 1992 22:18 UTC

Return-Path: <owner-chassismib>
Received: by CS.UTK.EDU (5.61++/2.8s-UTK) id AA16233; Wed, 12 Aug 92 18:18:00 -0400
Received: from relay1.UU.NET by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA16229; Wed, 12 Aug 92 18:17:48 -0400
Received: from uunet.uu.net (via LOCALHOST.UU.NET) by relay1.UU.NET with SMTP (5.61/UUNET-internet-primary) id AA01522; Wed, 12 Aug 92 18:17:12 -0400
Received: from sam.UUCP by uunet.uu.net with UUCP/RMAIL (queueing-rmail) id 181605.6926; Wed, 12 Aug 1992 18:16:05 EDT
Received: by davidsys.com (DECUS UUCP ///1.3-1/2.5/); Wed, 12 Aug 92 15:01:29 PST
Date: Wed, 12 Aug 1992 15:01:29 -0800
Message-Id: <0095EFD88D6B94A0.21E004CB@davidsys.com>
From: bierman@davidsys.com
Subject: uptimes and timestamps
To: chassismib@cs.utk.edu
X-Vms-Mail-To: UUCP%"chassismib@cs.utk.edu"




The objects chasSlotLastChange and chasEntityTimeStamp are based on sysUpTime.
The problem is that sysUpTime is defined as the time since the "network
management portion" of the system was last re-initialized.  There are
devices (e.g. repeaters) that can operate without the presence of
a management entity.  Also, the hardware portion of a device may be
reset independently of the management portion.

We have been using a object called "chassisUpTime" in our agents for awhile,
indicating the time power has been applied to the chassis.  Objects such as
a chassisSlotTimeStamp are relative to the chassisUpTime.  This prevents
the situation in which a hardware uptime (or timestamp) has to be reset 
when sysUpTime is reset.

So I'm proposing this additional object:

          chasUpTime   OBJECT-TYPE
              SYNTAX    TimeTicks
              ACCESS    read-only
              STATUS    mandatory
              DESCRIPTION
                      "The time (in hundredths of a second) that
                      power has been applied to the chassis."
              ::= { chassis 5 }


I realize that many vendors have been using sysUpTime in this way (I believe
incorrectly). I would also propose that chasSlotLastChange and possibly
chasEntityTimeStamp be relative to chasUpTime, not sysUpTime.


--andy;
bierman@davidsys.com