I got a question this week from a customer about Unitrends upcoming deduplication capability (Adaptive Deduplication) in its release 5. The customer asked if we were going to support deduplication to tape. Here’s how this part of the conversation went:
Customer: Will adaptive deduplication support deduplication to tape?
Me: Excuse me?
Customer: You know – like CommVault and CA are doing – where you can save on tape media cost by deduplicating data so you can put more on the tape.
Me: Before I answer, let me ask you a few questions.
Me: Do you want to be able to restore your deduplicated data from the tape even if you lose the data on your backup appliance?
Customer: Of course – that’s why I’m using tape.
Me: Do you want to be able to restore your deduplicated data from tape even if your deduplication cache is corrupted?
Customer: What’s a deduplication cache?
Me: It’s a cache of the actual deduplicated blocks that exists either on the backup appliance or on one or more tapes.
Customer: Then of course I do.
Me: Do you want to be able to restore your deduplicated data from a single tape?
Me: Do you mind buying more disk storage so you can save on tapes?
Customer: Of course not – that would be stupid.
Me: Okay – thanks for bearing with me – one more question. You know tapes are risky – do you mind increasing the risk of not getting your data back from tape?
Customer: Of course I mind.
Me: Then the answer to your question is equally simple – of course we’re not doing deduplication to tape. That’s what companies do that put technology ahead of their customers – and their customer’s data.
I’ll have to admit – I thought when CommVault came out and made such a big deal out of deduplicating to tape that people would see through it pretty quickly. Take a look at this article over at SearchDataBackup.com – Does Data Deduplication to Tape Storage Make Sense? The answer of everyone they asked was pretty much the same – deduplication to tape doesn’t make sense. I particularly liked what one guy who was quoted in the article said:
“While Paulk noted that the price point may be good, ‘if you can’t get your data off, it could cost you your job,’ he said.”
It doesn’t get much more concise than that.