If the Centera is using the standard M (MD5) naming scheme (most likley) then you will get SIS on the Centera. It does take up a small bit of space as it will create new CDF - Content Descriptor File, to point to the Blob that contains the file that is single instanced. So you will have 1 CDF for the journal and 1 CDF for the mailbox. Each CDF is 27 bytes.
I would also be looking at having separate pools for each different kind of archiving if it's not too late;
EG ; Journaling Pool, Email Pool, FSA Pool ... etc as its easier to manage, view usage and control access.
SIS will still work between pools on the same Centera. It all comes down to if there is already a file on there with the same MD5# then it has the same Content Address and will just create the new CDF and point to the existing file (blob).