EMC SourceOne with Data Domain (VS) EMC SourceOne with ATMOS .. For 20,000+ mail boxes archiving !

We're looking to implement SourceOne with ATMOS in our environment, but we also have Data Domain as well.  The difference between the two platforms could fill volumes, so I'll be brief.

Data Domain (DD) relies on a deduplication and compression engine to achieve huge economies of scale on stored files.  So, for archiving, DD will do commonality factoring - it will only write a new file once.  If your users have very similar files, DD will track and only write unique files once, then compress and make excellent use of space.

ATMOS uses an object-based approach, storing files without a traditional filesystem.  ATMOS can scale to billions or more objects, and manage the creation and deletion of those objects with policy (keep XXX number of copies of a file for YYY period of time, then reduce the total number of copies to ZZZ).  ATMOS can use traditional compression or traditional deduplication as well, and achieve some additional space savings as well.

As far as the differences between compression/deduplication on the two, most people would agree that DD takes a more "aggressive" approach to deduplication and compression and devotes much of its CPU cycles to it.  ATMOS takes less "aggressive" approach, since it does this as well, but also has to manage replication services, the metadata layer and so on.  It can also be said that DD typically costs more per TB than ATMOS - which is pretty easy to expand - so you have to decide where you want to spend your dollars.

Here's a nice post on DD with SourceOne: SourceOne Email Archives on Data Domain?

Let us know if that helps!

Karl