sometimes we cannot :S .. I had the experiences of getting from amazon tech support "sorry but the there was a hardware degradation in that sector" (what ever that means) or that the server just was unreachable in the middle of the process and we urgently needed to re-launche a new server..
I ask this paranoic questions because of the type of projects we do. They require very hi availability for a very short time.. they last 2-3 days because of the nature of the project. so we need a very flexible way to re-lauch in bad scenarios.
any ways, the schema is understood