Hi Peter,
Microsoft's Exchange Server 2010 eliminated the concept of Storage Groups. They simplified the terminology such that now what we think of as a "storage group" can contain only 1 database and 1 corresponding log set.
In the NetApp world, we indeed had the capability to back up and restore multiple databases stored on a single LUN provided they were all in the same "storage group" and shared one log set.
Now that we can only have 1 database per log set in Exchange 2010 (what we used to call a storage group), we are indeed limited to one database per LUN. This is due to the fact that what we were doing in prior versions of exchange when restoring a database was:
1) Restore all databases on the LUN...
2) Replay the logs for each database in the storage group forward from a single coordinated log set across all those databases in the same storage group.
When using CCR in 2007, the same limitation (1 DB per storage group) existed that exists now in Exchange 2010. The restriction is really centered around keeping the continuous replication engine simple and VSS friendly, while ensuring that log replay is reliable on all copies. Exchange 2010 fully marries us to that concept wheras Exchage 2003-2007 still supported multiple DBs per LUN in the same storage group so long as you were not using any continuous replication (log shipping) technology.
So, the bad news is that you do need one LUN per database. The good news is that you can share a log LUN across many databases if LUN count is an issue (usually recommended to keep this to less than 5 log sets per LUN). Wherever possible, I recommend that you use massive databases (up to 2 TB is comfortably recommended by MS), and have 1 lun for each database plus 1 lun for each log. If you have to compromise, use multiple logs sets per LUN and you will still be in a supported config.
Regards,
-Todd