HP Virtualization Manager and HP Capacity Advisor: an introduction

After seeing these two graphs, the question arises, “How much hardware, if any, can I save by
consolidating these workloads?” The initial answer from looking only at the previous two graphs
could be “None.” The first workload requires seven CPUs and the second requires eight CPUs,
so the consolidated system must have 15 CPUs. However, Figure 18-3 shows the aggregation of
the two consolidated workloads. From this graph it is clear that while the peaks appear to line
up very closely, they are offset enough to provide an opportunity for hardware savings. In this
case, the sum of the two workloads is less than 12 CPUs. This is a 20% savings in hardware
without a reduction in the quality of service or response time.
From this simple experiment based on production workloads in HP’s internal IT depart-
ment, it is obvious that careful capacity planning can truly result in higher system utilization
and reduced hardware requirements; both will result in cost savings. The following section walks
through an example Capacity Advisor scenario that describes how the product can be used to
effectively and reliably plan for workload consolidations.
Capacity Advisor Example Scenario
This scenario involves three production workloads that are each running in separate nPartitions,
as shown in Figure 18-4. The goal of the scenario is to determine whether one of the nPartitions
can be eliminated and the resources of the other two nPartitions can be combined into a single
nPartition that will host all three workloads. This environment reduces the number of operat-
ing systems that require maintenance and licensing by two-thirds and reduces the hardware re-
quired by one-third. However, these savings must be weighed against the potential costs associated
with poorly performing workloads if the workloads peak at similar times or otherwise have re-
source contention during their normal operation.
chapter 18 Capacity Advisor
445
Figure 18-3 Combined Utilization for Both Workloads