It also depends on who you are talking to and what the context is. For
many people "utilization" is the average CPU usaged over longer
periods of time. Something you consider in capacity planning. This is
where you size the peak capacity of your LPAR based on the measured
avarage capacity.  In that aspect, VM is often different from MVS.

To drive a system that busy you need workload. When much of your
workload is through jobs that are scheduled in the way to optimize
throughput etc, then you may achieve a very high average utilization.
If most of your workload comes from end-users who are less easy to
control and end up wanting their share during the same office hours,
you find your average utilization much lower. This is often what
people refer to when they say VM maximum utilization is lower: the
typical VM workload is not as easy spread over the full 24 hours of
the day.

Now discrete servers have it even harder - their workload is for one
single application or end-user, and even more likely to have short
peaks and long periods of not usage at all.

Rob
--
Rob van der Heij
Velocity Software, Inc
http://velocitysoftware.com/

Reply via email to