To get equivalent performance, you need to define your measure of what
your performance is. Do you measure it by transactions per second? by
average response time experienced by the end-user? There may be others
that are important to you.

After you do that, though ... there's not much more "science" to be
had in this. You can't really predict how a given UNIX-based system is
going to perform. (You can do relative performance, and say "I would
expect this UNIX system to be 25-40% faster than this box") but
there's really no way to estimate how a particular system is going to
act until you've moved to it.

What you might want to do it rent or lease a system that you think
might work. If it meets your needs, buy it. If not, let it go and get
something bigger.

On Tue, Jun 3, 2008 at 10:27 AM, Brent Moore <[log in to unmask]> wrote:
> I have been putting this off but I think I need to start looking. I will stay with
> the HP hardware and was starting to look into the 9000 box but with the
> announcement I think maybe the Integrity systems. I currently have an A500-
> 200MHz-2 way system, 1024MB memory, 18GB system disc, 54GB user disc in
> an array, DAT24 and DLT80. I set my limits at 200 sessions and 15 jobs and
> average 145 combined. I use Image and Speedware and all systems are
> written internal so moving to a 9000 box seemed to be cost effective. I want
> at least the same performance or better. Has anyone out there moved from
> the 3000 A box to the 9000 box or Integrity and got any suggestions on
> models. ie rp2450 or rx3600. Thanks
>
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
>



-- 
"If hard work were such a wonderful thing, you'd think rich people
would have kept it all for themselves" - Lane Kirkland

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *