All you Cloud aficionados out there probably already know that Amazon Web Services (AWS) offers multiple availability zones (AZs) in multiple data centers around the world. The idea of an AZ is to provide a measure of redundancy and fault tolerance across the Amazon Cloud. In theory, a problem could take down one AZ but leave others standing.
Given they have multiple AZs scattered across many data centers, just how many Clouds does Amazon have, anyway? In a fundamental sense, they have only one. AWS customers can use a single interface to provision, manage, and deprovision instances across all AZs. Furthermore, you can use that same interface to set up load balancing and failover across AZs. In a very real sense, therefore, all of AWS’s AZs roll up into a single Cloud.
There’s no guarantee that other Cloud Service Providers (CSPs) offer the same seamless cross-data center management and provisioning. I won’t point fingers, because the CSP marketplace is so dynamic, but take a closer look at the likes of Rackspace and Verizon Terremark, as well as Cloud offerings from IBM, HP, and Oracle. Yes, they may have several data centers. Yes, they may even have AZs. But do they have a single management and provisioning interface that customers can use to provision Cloud instances across their entire infrastructure?
You might think that more Clouds are better. But in this case, CSPs should offer a single Cloud each. If they offer more than one, perhaps they haven’t worked the kinks out of their Cloud offering yet. Be warned!