In our last entry we discussed how the encapsulation of the millions of files that make up a single server into a single file has changed what we expect from storage and the network infrastructure. The same encapsulation may have an even more dramatic impact on data protection that will fundamentally change the way you protect and recover data. It may also change who the 'go to' data protection vendors are in the near future.The difference again as we discussed last entry is now an entire server can be protected by just copying one file to the backup destination instead of millions and millions of files. Further by integrating with the virtualization software via an API like VMware's vStorage API, these backup processes can be further improved. A good example is the ability to have only the changed blocks of a VMDK file sent to the backup application, reducing the amount of data that needs to be sent across the network.
As we discussed in our recent article "VMware Backup: Feature or Platform" the big change and possible challenge for traditional backup vendors is going to be what they do with this encapsulation from a recovery stand point. If all the backup vendor is doing is storing the backup image in their proprietary backup format, then your recovery process is going to be very similar to what it is today. Backup vendors need to re-think that strategy.
Modern backup solutions should expose these backups of encapsulated servers in a way that will allow you direct access to them. If this can it could be a game changer. For example you would be able to start a protected VM directly from the backup storage device, speeding the recovery effort substantially. While the VM might be in a degraded mode when running from the backup device that is better than having the application down while waiting for the data to move across the network. Then with the application back up and running, you could leverage a capability like storage vMotion to move the data to production storage in the background.
Further this type of access to the virtual machine should allow you to launch a secondary copy of the server for testing. You could even use this secondary copy to start an application like Exchange or Sharepoint and recover individual objects like email messages from those data sets. Doing so would eliminate the need for operating specific and application specific agents. This also then does away with the need for separate backup processes to get application level information.
Encapsulation of server data is often an overlooked benefit of server virtualization. The reduction of millions of files to a single file is fundamentally changing the way we manage data. How applications that manage data take advantage of this change is going to determine who the storage management leaders are in the future.
Track us on Twitter: http://twitter.com/storageswiss
Subscribe to our RSS feed.
George Crump is lead analyst of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. Find Storage Switzerland's disclosure statement here.