[This is the third and last excerpt from our Six Fairy Tales of VMware and Hyper-V Backup whitepaper.]
One of the more prominent fairy tales floating around is that Hyper-V host-level backup is efficient. Hyper-V host-level backup will be efficient one day – but it’s not now. The reason is the underlying technology that Microsoft uses to protect all of its operating systems and applications. This technology, which is known as VSS (Volume Shadow Copy Service), is also used to protect Hyper-V environments.
Unfortunately, VSS for Hyper-V at the host-level only supports full backups at the time this paper was being written. This means that protecting Hyper-V is currently better done by protecting each Hyper-V virtual machine as if it were a physical system, as well as protecting the underlying Hyper-V physical system using the same technique.
Another fairy tale that many people believe is that source-level deduplication is an effective technique to use with Hyper-V environments. A consequence of each VSS-based Hyper-V backup being a full backup is that if you’re using Hyper-V host-level backup, then you are generating a tremendous amount of redundant information. This makes sense, right? After all, the reason that modern backup systems don’t use full backups all the time is to perform data protection more efficiently by eliminating redundant data.
In order for the glass slipper of data protection to fit with respect to Hyper-V, some vendors recommend using source-level deduplication. That way, the network over which these backups have to travel and the storage used to store the backup will be less burdened. Of course, the problem is that by using source-level deduplication, you are consuming more resources on the Hyper-V physical computer, which lowers the ability of that computer to effectively host more virtual machines.
The answer is the same as described in the first fairy tale: protect all Hyper-V physical and virtual machines as if they were physical machines.