Re: snmpv2 pros and cons

Bill Norton <wbn@merit.edu> Tue, 12 October 1993 01:04 UTC

Message-Id: <9310112230.AA11644@merit.edu>
To: snmp@psi.com
Subject: Re: snmpv2 pros and cons
Date: Mon, 11 Oct 1993 18:30:51 -0400
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Bill Norton <wbn@merit.edu>

Hmm...

  >I must have been sleeping during earlier exchanges (:-(); ISODE is
  >manageable through SNMP2? What is going on here? SNMP1? CMIP? CMOT? (Sorry
  >to be so dense.)

I am refering to the ISODE snmpv2 code.  There is a separate snmpv2
distribution containing an SNMPv2 agent to build on top of the ISODE 8.0
package.  And no, it doesn't instrument all of the stuff in the ISODE package.
...and be careful about using those four letter words in this forum ;-)

  >Have there not been proposals for these problems?

Sure - the NMS will take care of it (gulp).

Note that these (clock skewing necessitating resynching, etc.) are *potential*
snaffus - ones that will be difficult to detect/recreate/diagnose/etc...
The problems with SNMPv1 implementations that Karl has found will likely
continue to exist, but to a  greater degree because of the increased 
complexity in V2 implementations. ( As the complexity of the spec
increases,  so does the chance for someone to screw something up, right? ) The
difference is with V2 the data is potentially encrypted and containing
time-sensitive data.  

BTW I *really* rather hear about real experiences with this stuff as 
opposed to postulation.  Any V2 developers / deployers care to
contribute their experiences to the group?

Bill