Is Data Protection Improving?

Disk-based backup (D2D) providers constantly ask me for one stat: What percentage of backup or recovery operations from tape fail?

I really hate answering this question. Not because I don't have a stat handy, but because it's a loaded question and the "answer" is really irrelevant. After all, how many end-users do you know who actually monitor the success and/or failure rates of their backup and recovery operations? And the answer doesn't matter if you're talking disk or tape backup.

In fact, we estimate that only 2% of end-users are, even though the reporting technologies have been around for quite sometime. Standalone reporting tools are available from companies like Bocada and WysDM, and many backup and recovery software providers enable basic monitoring and reporting.

So, let's face it, vendors that ask this question are really just looking for more fuel to fire the disk-versus-tape backup debate, aren't they? Absolutely. But to these vendors, I ask: What tape vs. disk debate? End-users have known for years that their traditional backup and recovery methods are grossly inadequate. They know that if they need to recover data from tape in a recovery or outage situation, well, frankly, they're likely in big doo doo. So, again, what difference does it make if you say that 20% or 30% or 60% of all backup and recovery operations from tape fail?

For argument's sake, let's just say that the failure rate is only 5%. Sounds good, right? No, not if mission-critical data is left unprotected. And that's why technologies like CDP and near-CDP are so important. With CDP, it doesn't matter what the failure rate is because you're continually capturing changes so there shouldn't be any data loss.

So, why play this statistical card at all? ESG Research shows that end-users are ready to make significant changes to their data protection environments. They are implementing disk-based backup technologies such as VTL and near-line disk in droves. Why? Not because disk is any more "reliable" than tape, but because it is a heck of a lot more efficient for recovery.

A better question might be the following: How quickly can you recover data in a failure situation?