At first blush, it sounds like a mystic’s riddle, somewhere in the vicinity of the sound of one hand clapping. But in reality, it’s a setup that makes perfect sense: Rackspace runs its OpenStack cloud in another OpenStack cloud.
“We’re running our cloud inside of a cloud,” John Engates, chief technology officer at Rackspace, recently told Wired.com. “All of the control nodes that are necessary to serve the customers on our cloud are running in another OpenStack cloud.”
The arrangement is only natural. “It’s a better way of doing things,” Engates said. “We’re merely eating our own dog food. The application we’re offering to customers is running on that same application.”
OpenStack is a way of pooling resources from a vast collection of machines, including processing power and storage space. Rather than running your software application on a particular server, you run it on OpenStack, a platform that spans hundreds of servers, and this platform can grab you as much processing power as you need, whenever you need it. This makes it easier to launch applications, but it also makes it easier to expand or “scale” them – to reach more users with more servers. And when a server fails, the platform is smart enough to move the machine’s work to a new one.