2. I've heard of hundreds, even thousands of rules. SWmetering is one of
the quietest things on the client. If there's a box that might get a big
hit.. I would say it would be a Terminal server; where potentially a few
hundred users are hitting the same box. Those you might want to watch their
logs; and perhaps decide to use a Local Override Policy to either complete
disable the SWMetering Agent on those clients, or set different frequencies
of summarization (more often) so it's less to report per instance.
3. It watches everything; but only tags things for which rules are defined.
Again--it's very quiet. I wouldn't worry about it.
4. The eternal answer: it depends. It depends on your environment. I was
at a financial company for a while; and there were certain applications that
were only used once a year during tax season--but the users needed those
applications to be there. We had to get a full years' worth of data before
enough data was available to definitively say Computer Y didn't need
Application X. But for something really expensive per license (like
AutoCad), 2 months of data would be enough for most license reclamation
project managers.
"net1994" wrote:
> We have never utilized Software Metering before in our SMS 2003 environment
> and we would like to start. A few questions:
>
> 1, I have a feeling once metering is enabled on the client, it will only
> report usage from that day onward? So we cannot tell how many times iTunes
> was started before software metering was enabled. Right?
>
> 2. If I create 7 Software meter rules for clients to report back on, will
> this kill the client resource wise? How will it affect CPU, and RAM when the
> data collection is running?
>
> 3. Does data collection process scan every process all the time on a pc, or
> only wait for a metered program to start?
>
> 3. What is a good rule of thumb to wait after metering has been enabled
> before we can get accurate data-14, 30, or 60 days? I'm interested mostly in
> programs installed, but not used.
>
> Thanks