I'm a programmer and what goes on in the data centre is a mystery to me, so this question might be very very stupid.

If you have multiple backups of the same data, and you "de-dupe" it, doesn't that mean you now only have one backup of that piece of data, so only one point of failure?

How does that help? Surely if you're running full backups you want discrete copies rather than de-duped copies, which are more-or-less equivalent to an incremental backup?

And, conversely, if you want de-duped backup data you just run an incremental backup in the first place, without paying anyone a dime for de-dupe techonology.

What am I missing here? De-dupe sounds like a stupid idea for backup data.

Paris, because I wanted an icon with a question mark on it, and I can see the IT angle.

"If you have multiple backups of the same data, and you "de-dupe" it, doesn't that mean you now only have one backup of that piece of data, so only one point of failure?" ... By Eddie Edwards Posted Monday 3rd November 2008 16:13 GMT

Eddie, You also have one point of Source thus is all Progress of Mutual Mind and Mutually MetaDataMined. Which is QuITe Clever.

Can I has Fuse?

I used to think I was... not stupid... until I read this article. Is this like non lossy JPG? surely squashing the data down further and further until it reaches a quantum singularity will drag everyones' data into a binary black hole. I can see the Reg headline now:

Cloud computing disappears up its own backside.

That's it, I'm off to rebuild a PC (faster, stronger, better - we can rebuild you), and make myself feel better.

Eddie,

"And, conversely, if you want de-duped backup data you just run an incremental backup in the first place, without paying anyone a dime for de-dupe techonology."

where you're backing-up dozens or hundreds or thousands of machines/VMs including their OS system files, that still leaves a lot of duplicated data.

I would imagine you'd still want to keep multiple backups, maybe on different sets of disks/different locations..