Because of the massive amount of change data we have per day we don’t do any deduplication on our data. We have about 3 TB of file data with approx. 6-8% of the data changed daily.
perfect answer for you might be de-deduplication at the target versus in-band. In that case, your data at rest would have a smaller footprint which works hand in hand with thinly provisioned luns.
Our data is not that fluid or dynamic like Mr. Mrdenny. We are initiating a project to implement de-duplication. There are claims of 1 to 10; to 1 to 50 reductions of de-duplication data. The technology is mature and most leading vendors are engaged in this space. If you store Personally Identifiable Information (PII) then you will need to encrypt the data due to federal regulations. De-Duplication does provide an option to encrypt the data. I recommend exploring de-duplication arena for various options.