The Myth of High Availability as a Back-Up Solution

Restore from backup has always been the last-resort fail-safe.  However, many cloud-based services don’t offer that safety net. If you were to ask your cloud service provider how you can get a backup of your data, they will frequently sidestep the question. Their answer is sort of a magician’s sleight-of-hand trick:  “You don’t need a backup because you have high availability.”
But that’s a dangerous misperception. Many cloud providers take the position that so long as their customers have high availability, they don’t need backups. Even Microsoftpromotes this view, and adds that backup versions of data can be stored in the One Drive for Business (formerly SkyDrive) cloud storage that comes with an Office 365 subscription. All cloud-stored data should be synced in real-time over to multiple servers, the argument goes, so the client doesn’t need back-ups. This attitude has spawned an entire category of third-party software designed specifically to provide cloud-based backup solutions.
So what’s the difference between a backup and high availability? Basically, it’s a question of timing.
It may be accurate to say that high availability provides a sort of real-time backup. But what you lose in that scenario is the “snapshot in time” functionality that a backup system provides.  With high availability, there are multiple places where one can look at a current data set, but nowhere to see what it looked like a month ago. This can become a serious issue, especially if you’re in a highly regulated industry that requires a complete and detailed paper trail, such as healthcare (HIPAA), finance (SEC) or the biopharmaceutical and food industries (FDA).Microsoft’s Office 365 Trust Center is an excellent resource for researching how its software complies with these requirements.
There are at least two things backups protect your data from that high availability does not.
The first is file corruption.  On a high-availability system, if you have a file that becomes corrupted – a document, a spreadsheet, an email or anything else – that corrupted file is perfectly replicated, including the corruption, to all of the high-availability servers. It’s now perfectly replicated in its corrupted state in every data center. How is that helpful to you? It’s not.
The second issue is a deleted file. With high availability as your backup, the deleted file is completely gone. It can’t be recovered because it has been deleted everywhere. So there’s no protection against accidental deletions. Sure, there are other safety nets that exist, such as deleted items, but those elements are temporary and the amount of data saved can be dependent on several factors, such as storage space and time limitations.
In addition to the potential loss of data, high availability also does not give you a history of your configuration settings.
Our view is that it’s important to have backups in place that minimize loss of data, and that meet any regulatory requirements with rules covering the retention of emails and other documents. Make sure you know how your cloud provider will handle backups and be prepared to sacrifice a safety net, or come up with an add-on solution (i.e., via third-party software designed specifically for that purpose, or custom-written PowerShell scripts). Don’t accept high availability as the only answer when you move to the cloud. Know that high availability is not intended to be the same thing as backup.

Leave a Reply

Your email address will not be published. Required fields are marked *