Virtualization puts up Capacity Management challenges that have not been experienced for some time. After many years of running single application servers with hardware so over-specified that peak utilization has rarely gone over 20%, we now face putting many applications onto common hardware. Avoiding issues of contention for shared resources therefore becomes critical in ensuring service levels continue to be met. Even in a VMware environment high utilization levels can only be sensibly and safely achieved with mixed workload types. Disparate and heterogeneous farms of applications servers can be consolidated to a single homogenous pile of blade servers but all the operating systems and applications that existed before the exercise still exist and require management.

New dimensions are added to the performance manager’s task. As a new environment, many users are unsure of what overheads to allow for from VMware itself and how to have an adaptable Capacity Management framework that will enable you to identify problems before they arise and manage your systems proactively so problems do not take hold.

Many organizations are offering basic ‘add up the utilizations’ services to help predict what virtualized resources you need. This ignores the effect of many workloads competing for the same finite resources however. It also fails to take account of changes in workloads over time after virtualization due to business changes. Effective Capacity Management will overcome these limitations to ensure a smooth transition and the provision of a service aligned to business needs as your virtual systems mature. Good Capacity Management implemented to ITIL standards will be the basis of successful virtualization implementations.