RE: Call for censensus on path forward

                                                                                                               
                                                                                                               
                                                                                                               


> Also, I feel (but this
> needs to be verified), that Wes's proposal means substantial changes
> in all management information access code. This is just a show stopper.
>
> On the other hand, I believe that the GETCOLs proposal might be
> implementable with changes only to the SNMP agent.

Interesting.  We're interested in Wes's proposal precisely because we
believe that it will be implementable with changes only to the SNMP agent.
The agent will have to work at it a bit, but isn't this the scenario:

  - PDU: "Get columns { 4,5,6 } from table foo for those entries where the
value in column 8 = 300"
  - Agent to subagent or instrumentation: For each row i , Get (foo.4.i,
foo.5.i, foo.6.i, foo.8.i)

Then the agent looks at the column 8 values it retrieves, to decide which
rows to use and which ones to discard.

If this is the way it works, I don't see how the subagent/instrumentation
is affected at all, or even knows that anything has changed.

Regards,
Bob

Bob Moore
Advanced Design and Technology
Application Integration Middleware Division
IBM Software Group
+1-919-254-4436
[email protected]