Because this model assigns settings to both a resource pool and the virtual machines, any updates require a shutdown and restart of the virtual machines. The percentage guarantee for memory is applied to every virtual machine in this virtual datacenter, similar to the Pay As You Go model. Additionally, the settings are applied to the resource pool in vSphere like the Reservation Pool model.

With the settings shown in Figure 6, 100% memory reservation is available on the vSphere resource pool, and 50% of the CPU reservation is available. Each virtual machine created also requires a 100% virtual machine memory reservation within the pool to run. After all pool resources are assigned, no more virtual machines are allowed to run, but the settings can be adjusted. Because the resource pool expandable reservation is disabled, in most cases VMware admission control will prevent too many virtual machines from being powered on.

Lowering the percentage guarantee below 75% or even 50% can also create challenges as the resource pool gets only a fraction of the configured resources shown as Available Reservations. Virtual machines are still required to reserve those same values at power on. When using the Allocation Pool model, higher values are always better to prevent any limitations within the organization virtual datacenter.

The Allocation Pool model can also leverage elastic provider virtual datacenters in vCloud Director. The Allocation Pool model can use this added capacity automatically, without the affecting Company2. They can continue to deploy virtual machines as long as there is capacity to meet their requirements.

Understanding the effect of these settings on vSphere is crucial to implementation.

As with all of the allocation models, the provider needs to monitor the customer’s virtual datacenters and keep the customer informed about resource usage. If a virtual datacenter runs out of resources, vCloud Director sends an error message to the customer.