The design of a virtual infrastructure that will adequately address enterprise needs, like so much else in business, operates according to the laws of supply and demand. In the business world, these market forces help to determine the quantities of a product that will be both sought by consumers and available to them. A similar set of phenomena are at work in a virtualized computing environment, but in this case, the product being sought is the allocation of physical resources and the ‘consumers’ seeking those resources are the virtual machines themselves.
In order to provide as much computing power for as little an expenditure as possible, most administrators of virtualized environments want to assign to each physical host the maximum number of virtual machines that it can possibly handle. However, such administrators must also keep in mind the limitations of such an approach; too many VMs crammed onto the same physical host will result in slower performance for employees relying on the virtualized computing environment.
Capacity management therefore becomes an issue of give-and-take, with IT services personnel trying to balance the pressure for resources with a realistic allocation that enables optimal performance. Enterprises running VMware vSphere as part of a managed services model are able to harvest supply and demand information from the management console in the software to help them utilize resources in the best possible configuration. These data are known as "metrics" in the parlance of virtualization and can assist the IT department in their efforts to manage capacity in a way that enhances enterprise workflow.