Selecting a machine that will perform conversion

Take into account the following considerations.

Which agent is installed on the machine?

The resulting virtual machine type and location depend on the agent that resides on the selected machine.

What is the machine's processing power?

Conversion will take the selected machine's CPU resource. Multiple conversion tasks will be queued on that machine and it may take considerable time to complete them all. Consider this when creating a centralized backup plan with conversion for multiple machines or multiple local backup plans using the same machine for conversion.

What storage will be used for the virtual machines?

Network usage

As opposed to ordinary backups (TIB files), virtual machine files are transferred uncompressed through the network. Therefore, using a SAN or a storage local to the machine that performs conversion is the best choice from the network usage standpoint. A local disk is not an option though, if the conversion is performed by the same machine that is backed up. Using a NAS also makes good sense.

Storage space

For VMware, Hyper-V and Virtual PC, disks of the resulting virtual machine will use as much storage space as the original data occupies. Assuming that the original disk size is 100 GB and the disk stores 10 GB of data, the corresponding virtual disk will occupy about 10 GB. VMware calls this format "thin provisioning", Microsoft uses the "dynamically expanding disk" term. Since the space is not pre-allocated, the physical storage is expected to have sufficient free space for the virtual disks to increase in size.

For KVM or RHEV, disks of the resulting virtual machine will have the raw format. This means that virtual disk size is always equal to the original disk capacity. Assuming that the original disk size is 100 GB, the corresponding virtual disk will occupy 100 GB even if the disk stores 10 GB of data.