Nice to see a discussion getting started, and I really appreciate all your ideas.

Just talked to management, and before we start looking at geo-clusters, mirroring at disk level (really like that idea), as always, I need to try and figure out what is possible without any of that.

We all agree though, seeing that network bandwidth would be our bottleneck, that we should definitely invest some in that area.

Btw, db sizes range from a few MBs to a several dozen GB, with most of them being around 5GB.

I would also add that there's no considerable growth observed during the last year, so no problems there.

To add one more level of complexity, we have 2 production clusters with 500 DBs EACH, and we would like to DR that to another location but to ONE Server if possible (mind you this would really just serve the disaster scenario, so the server probably won't be as powerful as those on Production) ?

Do you think this changes anything somehow?

Thanks,