|
If I were writing the technical response for their RFC, I'd be advocating nightly backups of the server re it's structure and then I'd have a Network storage device clustered with one of the cluster devices stashed at our isp.
If the machine dies, you replace it. Quickly. Just have a spare or a suitable replacement and bang the Server State backup onto it.
With the mirrored storage, if your local goes down, you use the remote at a lower latency, but you're still live. Rebuild and restore the local.
You pay only for the synch data, which is a lot less.
|