One of the things I'm investigating is DR, actually not really disaster recovery, more of a "service availability" problem. I'm not trying to address the DR situation yet, just the "OMG SERVER CRASH" problem.
So I started doing research on clustering vs mirroring, since I hadn't done it before. If you're researching mirroring, start here:
That's a technet whitepaper on mirroring. After you're done with that, go here:
Performance implications of mirroring, along with best practices stuff. The interesting part of the second paper is the performance implications of the encryption options (negligible impact, should always be turned on). I found...