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

Re: Protocol operations proposal deadline



HI,

I have a fundamental  problem of viewing filtering as a way to "remove"
the "columns" (attributes) returned by an operation. This implies that
the operation is retrieving all columns (attributes) of all instances
(rows) of an object class (table). This is both wasteful and limiting.

More below....
At 08:11 AM 9/13/2002 -0700, Wes Hardaker wrote:
>>>>>> On Tue, 10 Sep 2002 07:36:24 -0400, Mark Ellison <ellison@ieee.org> said:
>
>Mark> (3) I like very much the fundamental suggestions in
>Mark> draft-hardaker-eos-oops.
>...
>Mark> I would also like to see additional filtering added to the new
>Mark> oops operations.  There are two sorts of filtering to suggest
>Mark> here.  The first is a mechanism to specify which components
>Mark> within an oo-object should be returned (e.g.  which columns
>Mark> within a row).
>
>Mark,
>
>FYI, I asked this exact question in Yokohama and got an overwhelming
>"yes", so there is support already for it in the next version of the
>draft which I intend to post sometime in the next few days.
>
>Mark> The second sort of filtering would require the CG to supply a
>Mark> secondary varbind list containing metadata (in a proper
>Mark> placement within the oops message) to facilitate getting the
>Mark> next object/row with a particular characteristic.  An example
>Mark> was posted wrt to scli where the next ifEntry associated with an
>Mark> ethernet interface was desired.  This mechanism might prove
>Mark> useful in reducing overall message exchange between CG and CR.
>
>There are more filtering changes in the next version with respect to
>requesting a particular subset of data as well.  Take a look at it
>(next week) and let me know what you think.
Now, it turns out that I do like filtering, but view it as a mechanism
that a management app writer can use to be very wasteful and blame
the agent implementation in the managed system.

Note that filtering can be based on values of instance identification,
as well as the values of attributes, and higher level values such
as "return columns a,b, and c of this table, where the value of
column d is greater than one standard deviation of the mean value
of column d." Such a request can be quite useful or wasteful.
 
Regards,
/david t. perkins