I'm trying to understand how deduplication works, I've read some artticles and still cant really understand it.
If I have say 1TB of data on a server and I back that up to disk using deduplication then I have 1TB of data. The next day I do an incremental and the backup job report report says 90% was deduped (only 10% changed) so it only backed up that 10% which changed. What I cant understand is,isnt that what an incremental backup does anyway even without using deduplication? It only backs up what has changed.
If we continue with that scenario that a full backup on Sunday 1TB. Mon-Sat I run an incremental and each day only 10% of the data has changed, then I have backed up in total 1TB (Full) plus an extra 600GB (100GB per day) for the incremental Mon-Sat - so it backed up in 1 week 1.6TB of data.
When I did incremental backups to tape, it was the same thing, so whats the difference?
Or (as I just realized writing this), am I totally missing the point and deduplication only means that if for example there are 50 unique files scattered on a server (or servers) then deduplication only takes one copy of that file and skips the other 49, meaning that it's saving you diskspace where the backups are being stored?
Do you mostly use client side deduplication or deduplication on the media server?
Thanks!
