Canon Information Systems
(714) 438-7163 - voice
(714) 432-1531 - fax
>From: JK Martin[SMTP:firstname.lastname@example.org]
>Sent: Monday, April 28, 1997 12:43 PM
>To: Farrell, Lee
>Subject: Re: PMP> DMTF / CIM -- Printer vendor participation?
>Thanks for the heads-up on this new DMTF work.
>Did they say anything about the WBEM initiative announced way back last
>July? Seems to me that the WBEM info announced by Microsoft alluded to
>some sort of "partnership" of sorts with the DMTF, such that they were
>going to share in the development of WBEM (in some way).
Yes, there was much reference to WBEM. It seems that CIM is just one
component of Microsoft's grand vision for WBEM. (Originally, it was
called HMMS -- but the DMTF took it over and made it "better." Now
Microsoft endorses CIM.) Not surprisingly, Intel (which seems to be
driving most of DMTF) is also a strong supporter of the WBEM initiative.
More info on WBEM is available at http://wbem.freerange.com/
>Also, was Sun's new Java-based "JMAPI" work discussed for integration?
JMAPI, CORBA, HMM and "Other" were all referenced as alternatives for
exchanging management objects/information. DMTF claims that all will be
able to take advantage of the CIM model (and vice-versa). However, I
didn't hear any specifics about JMAPI per se. Java (and COM and C++)
were referenced as possible options for interfacing between a Management
application and Microsoft's Hypermedia Object Manager (HMOM) -- another
of the WBEM entities. Evidently, Novell's ManageWise is planning on (or
already has?) some sort of Java-based Management approach.
>How about any references to the various CORBA-related activities going
>on with net-based management?
Again, no real specifics -- just that the CIM model would be useful to
it. (I can't swear to it, but I got the impression that Tivoli/IBM is
big on CORBA? Tivoli is a major impetus behind the CIM.)
>Whatever the DMTF comes up with will certainly be interesting to
>review. However, it's hard to imagine that they are doing anything
>more than simply reinventing the wheel for the bulk of their effort.
>What does their approach offer the world that doesn't already exist in
>the other approaches? Where is the differentiation? Why would they
>roll their own, rather than joining an existing effort?
If I absorbed the message correctly, I think the CIM model is an attempt
to unify (and offer more capability beyond) the current implementations
of SNMP, CMIP, DMI and "Other." It is more of a focus for enabling
Management applications to be platform and/or device management
implementation independent. The plan is for individual object providers
to act between the different implementations and the HMOM (via HMMP?).
>Lee, I don't expect you to answer most of these questions, but I was
>others might be able to shed some light here.
I'd like to hear other insights, too.