By Jon Stokes
May 08, 2008 - 05:05PM CT

This past Monday, managed hosting company RackSpace announced that a subsidiary of theirs, Mosso, is currently beta testing a "cloud storage" service that will compete with Amazon's S3. Priced identically with S3 ($0.15/GB per month), Mosso's CloudFS will add another line to the ever-expanding menu of ways that users can keep files on remote, distributed, redundant storage.

What the announcement and subsequent press coverage didn't highlight, however, is how Mosso actually got into the cloud storage game in the first place. For a managed hosting provider that's already in the business of building, deploying, and managing servers, rolling out a cloud storage solution wasn't a hardware problem as much as it was a software problem. A RackLabs blog post tells the tale of how the company conceived of and implemented CloudFS.

The story is almost as interesting for what it does not involve as for what it does: there are no huge storage area network (SAN) devices. "Rackspace is a hosting company," John Engates explained in the post. "We have a ton of perfectly good older servers around that are not in use any more. First of all you take some of those 'seasoned' servers and you rebuild them from the ground up with exactly what you need. Heck we've got the parts. We don't need lots of processing power. We don't need much RAM. We need disk space and a lot of it."

He went on to discuss the company's new provisioning system: "The servers are easy to build, and we do that hundreds of times a day for our other customers. CloudFS servers can be slotted anywhere in the DC and they boot off the network. If one fails, we just stick another one in its place. Fast and easy." In other words, rolling out the hardware side of CloudFS required Rackspace to do exactly what it was already doing anyway, with hardware that it already had on hand.