The VM File Copy Problem
VHD files tend to be very large and here are a few reasons why

• Microsoft recommends using fixed size VHD files
• Running out of disk space inside a VHD can be disastrous, so tendency is to err on the side of caution
• Unlike the physical world, VSS differencing area needs to be inside the same VHD

Unfortunately, many of the usage scenarios for Virtual Machines directly lead to a file copy. For example:

• A typical backup scenario involves creating a diskdshadow based VSS snasphot of a running VM, mounting the snapped volume as a read only volume, and copying the VHDs from this mounted volume
• DPM 2010 requires a full backup to start off - typically the full backup is done as a diskshadow based VM snapshot and then a xcopy or robocopy of the snapped volume containing the VM VHDs - a VHD file copy!
• Migrating from Hyper-V 1.0 based on Windows Server 2008 to Hyper-V 2.0 based on Windows Server 2008 R2 and its Clustered Shared Volumes (CSV) also leads to a file copy. If you do a SAN LUN transfer, the recommened 4 to 8 VHDs per CSV LUN will all move with the LUN transfer from SCVMM Library to a Hyper-V host.
• Moving VMs from a test server to a production server is a file copy operation
• Checking in or checking out a file to/from SCVMM is typically a file copy operation

The current solution - an inefficient file copy
For non running VMs, the current file copy operations are inefficient in a number of ways:

• Data that is unused by the file system inside the VM is copied
• Data that is used by the file system inside the VM, but will never be used is copied. For example, when a master VM is checked out, the hibernation file will also be copied. The checked out VMs will typically always be booted afresh and never resumed, at least the first time the VM is fired up
• Meaningful data is copied, and then immediately overwritten. For example, the current usage requires an admin to
1. Wait for a check out file copy operation to complete - and wait for unused parts of VHD to be copied
2. Boot the VM
3. Install some new software e.g. a SP or Microsoft SQL Server and this will overwrite a fair amount of the VHD file copied in step 1

Introducing VHDCopEE
The VMUtil Inc. family of products work together to remove the inefficiencies. The patent pending VHDCopEE refrains from copying meaningless portions of a VHD file. VHDCopEE and VMProv work jointly to minimize copying data and then immediately overwriting it.

VHDCopEE can accelerate fixed size VHD file copies.

• The source VHD can be on a local volume, a VSS snapshot read only volume, a network volume, or a CSV volume.
• The destination VHD can be on a local volume, a network volume, or a CSV volume.