Hyper-V Scalability Evident in Its New Virtual Disk Format
Microsoft’s Hyper-V 2012 includes a number of new features that improve its scalability. Among the most exciting of these features is the new virtual disk format, called VHDX. Here are some of the interesting features I’ve seen while looking at Hyper-V 2012 as well as reading through VHDX technical specifications.
- VHDX supports a maximum disk size of 64 TB, whereas the limit on the older, VHD format, was 2TB.
- Disk writes are first logged to an intent log in a metadata area prior to changes being written to their final disk location. This provides increased protection from data corruption in case of power failure. After recovery from power-failure, any changes that were in progress will be in the intent log and changes will be replayed from these logs.
- The new disk format supports a large 4KB logical sector size, and block sizes for virtual disks up to 256 MB; this allows for tuning of the block size to match the disk write patterns of the applications running in the VM.
- VHDX includes support for trim/unmap operations. These operations improve the efficiency of VHDX disks by allowing the physical storage to reclaim freed space (provided the host hardware and the guest OS support these operations).
- The disk file supports custom metadata, so that you can store information such as the system version or patches applied.
- You can revert VHDX files to the prior format easily by using the Hyper-V manager, editing the disk, and converting to a VHD.
Along with the many other improvements in Hyper-V 2012, this new virtual disk format will cause users who never considered Hyper-V as a viable alternative to step back and take notice. At Unitrends, we have placed a lot of emphasis on the protection of both virtual and physical environments. We have the Hyper-V 2012 release candidate running throughout our labs, so you can be sure that once this new virtualization platform is released, you will be able to fully protect your virtual environment with our solution.