Data is a fragile thing and poses a major Achilles heel to any business. Consequently, we're always looking for ways to better safeguard data. One idea that's been gaining considerable traction lately is cloud computing - the placing of processes and data onto the Internet. This means that a disaster at your site won't affect your IT operation and you can quickly resume business if you can gain Internet access.
However, as with any new technology, we still don't fully understand the risks of cloud computing and may well be exchanging one set of problems for another. This morning Amazon's Elastic Cloud Compute (EC2) service failed, bringing down a whole lot of businesses and services that use EC2. Apparently, EC2 is run out of a single data center in Virginia.
The lesson here is that you need to know whether your backup plan has its own backup. Have you just created a situation where you now have to worry about disasters in two locations? You should demand at least the same level of backup at the vendor site that you would have put in place at your site. After all, trying to mitigate disaster by adding a single point of failure to your backup plan doesn't really make sense, does it?