[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

MO Aggregation MIB.



Hi,
    I initially posted this mail to the disman list. Randy referred
me to this list. I looked at the charter and fully agreed. Thanks
Randy!.
    OK. So now let me get on with the story :-)

    We have been working on an aggregate MIB for some
time. I submitted the draft - but it did not make it before
the deadline. The draft is available on
http://www.cysol.co.jp/research/draft/aggrmib/index.html

    The essence is - the Managed Object Aggregation MIB
allows Managers (mid-level, delegates) to define Simple
Aggregate Managed Objects (SAMOs) in terms of MOs.
Accessing SAMOs instead of  the constituent MOs leads to
payload compression as the PDUs will contain one OID  -
that of the SAMO - instead of  all the OIDs of the constituent
MOs. The Aggregation can be extended to cover polling for
for high resolution, at intervals of the order of millisecs. In this
case an aggregate object SAMO can be defined to have a
value that is the concatenation of values of  the MOs at
regular intervals starting at a well defined time. 
    I have been following the other approaches at OID 
compression - (Juergen's and Sandy's) those are excellent
techniques.  I believe that this approach complements 
those . 
    Conceptually, I think that MIBs as they are now have been 
defined by equipment manufacturers and protocol developers
[without the specific purposes of management applications 
in mind.] Management application developers  should be able 
to define/configure MOs that serve their specific needs and 
effeciently. The aggregate MIB is attempting to make that possible,
without breaking anything (hopefully).
    I would like to start a discussion on this list and obtain
comments. If there is interest disman chair finds it
appropriate then I would be glad to make a 5 minute
presentation at the meeting in Salt Lake City (Is there one).
   Cheers

    Glenn