When the cloud is not really a cloud?

In reading this article on Information Week about the Amazon EC2 North Virginia outage, the first thing that came to mind is disappointment. Disappointment that the cloud turns out to be just another data centre. We have quickly come to expect that we should be insulated from underlying hardware issues. Putting software in the cloud should mean that it continues to operate, regardless of an issue with one piece of hardware. Redundancy and failover should be built into the system.

Apparently not. I guess this is a lesson for application providers. We cannot rely on the IAAS vendors, and will need to look at geography as part of the underlying risk assessment process. Perhaps most people already do this, and it's just me that had this expectation?