System Sizing Guidelines for Integrity Virtual Machines Deployment -- Hardware Consolidation with Integrity Virtual Machines

Preparing for Consolidation - Utilization Reconnaissance

When considering a workload for hardware consolidation, collect the utilization data that is

available, note its characteristics, and determine what information is important and what is not. If

you are preparing to consolidate several workloads, collect the utilization information (both

quantitative and qualitative) for every workload – not just for a few. You will need specific data to

determine which workloads can be located on the same physical server.

Where possible, collect utilization and capacity information for all basic hardware components,

including CPU, memory, network, and mass storage. Statistical utilization information for CPU is

usually available; detailed data for utilization of other resources is not always available. If that is

the case, collect qualitative information for the other resource utilization, especially I/O (network

and mass storage). Categorizing utilization of these resources as “high,” “medium,” or “low” will

be very useful for consolidation planning.

Peaks and Averages

Start consolidation planning with average utilization of the workloads. Then you can look at the

workload usage peaks to avoid consolidating workloads with coincident usage spikes on the same

hardware.

Consider spike or peak utilization characteristics of workloads that are being consolidating on the

same physical server. Random or out of phase peaks in utilization are not a consideration. Look

for workloads that are in phase with other workloads, for which peak utilization occurs

simultaneously. For example, if all the workloads are idle until the last five days of the month, then

average utilization is not useful for sizing considerations. For this kind of utilization pattern among

multiple workloads, use the peak utilization over the five peak utilization days in place of the

average utilization.