Q: Should I use Windows Server 2012's data deduplication if the back-end storage is on a SAN that's already doing data deduplication?
A: The Windows Server 2012 data deduplication feature works at a block level. It finds duplicate blocks of data within files, then stores that content once and links all the occurrences of that data from files to that central single instance storage of the data, removing the duplication of data and wasted space.
The process of looking for duplicate blocks happens by scanning the file system. This can happen in the background during times when the system is quiet, but it's not real-time, in that data is not deduplicated as it's written to the disk.
Many Storage Area Networks (SANs) and network storage appliances offer very similar capabilities that deduplicate at a block level, although SAN-based deduplication processes can be much faster at finding duplicate content.
The data deduplication by Windows Server 2012 and the deduplication done by the storage device will be unaware of each other. There would likely be very little additional savings having deduplication in the OS and on the storage appliance.
As data was deduplicated by Windows, the data wouldn't be duplicated on the SAN due to the single-instance storage. So, effectively, the data would be deduplicated only once. I would, therefore, probably only enable data duplication on either the Windows OS or the SAN, but not both. You might take a look at the FAQs "Q: How can I see details of Windows Server 2012 data deduplication storage?" and "Q: I have enabled data deduplication on a Windows Server 2012 volume, so why is nothing happening?"