Exchange deduplication rate is very low

With mailstore databases that are this small...

1) Why are you worrying?

2) I can say that NetBackup de-dupes at the 128 KB block level - so... the "dedupe stream handler" for the MS Exchange mailstore "data stream type" is going to be recieving about 41,000 segments - and if you receive only 5000 emails in one day - which, lets assume, land evenly across the 'mail boxes' then that represents 12% of possible dedupe-able segments which will appear to MSDP to be different...

...this is further compounded by the fact that a higher precentage (when compared to customers with much larger MS Exchange environments) of 'control and checksum and index and structure' blocks/areas within the mailstore databases will change - simply because the mailstore databases are so small.  So, I suggest that because your mailstore databases are so small that you therefore experience a higher rate of 'change density' within mailboxes but also a higher ratio of 'control structure change' per mailbox count... and so by the very nature of having such small mailstores, this means that even a 'low rate of emails' will generate a somewhat larger amount of dedupe difference, and thus a somewhat lower rate of de-dupe success.

This is of course all conjecture and unproven, and possiblly (probably?) nonsense, but does seem to be relatively plausible in my limited scope and understanding of such matters.