Capacity
planning
was
never
an
easy
task,
but
having
multiple
operating
systems
running
in
multiple
separate
partitions
creates
even
a
bigger
challenge
to
collect
all
necessary
performance
data
needed
for
estimating
performance
characteristics
and
applying
different
algorithms
for
different
operating
systems
goes
beyond
the
usual
"thumbs
in
the
air"
or
"adding
various
CPW
ratings"
with
pen
and
pencils
approach.
This
presentation
will
cover
important
aspects
of
sharing
resources
and
defining
upper
and
lower
limits
to
avoid
performance
impacts
across
partitions
and
still
optimizing
for
best
performance.
We
will
discuss
a
methodology
to
measure
performance
across
partitions
and
consolidating
estimates
while
recognizing
workload
peaks
and
time
zones.
|
