Thanks for the reply,

We are in the middle of moving to a new storage system to hopefully allow us to do block level replication to our DR location. We have had quite a hard time finding any solutions which will allow us to do any mirroring, as alot of technology was not written to quiesce 3000+ databases at once. Thats why we've ended up looking at the block level solutions. I'm not a fan of filter drivers but it seems like a necessary evil at this time.

We have a 24x7 global application but we do create our new user databases across different instances based on their timezone. This allows us to run maintenance jobs at times where it will least impact the customer. It also allows us to roll out updates to our application when users are not calling customer service because we disabled the system during the middle of their working day.

Has anyone else found a mirroring solution which works in an architecture like this? How do others in this scenario roll out schema changes when the application updates?