Last month, Rackspace and NASA jointly announced OpenStack.org, re-igniting a discussion about long-sought-after open cloud standards. We at Cloudkick believe strongly in cloud portability and standards; because there are currently no standards for cloud provider APIs, we created Libcloud to help with this issue. OpenStack is the first project that has a good chance at standardization between multiple cloud vendors and private clouds. So what makes OpenStack different from existing open cloud efforts?
There’s Interest Outside of the Software.
Both Rackspace and NASA have an interest in cloud, so much interest that when they couldn’t find what they wanted from a vendor, they had to invest directly and build it themselves. The technology behind the cloud was not the business interest; it was the services they can provide with it. For that reason, both NASA and Rackspace have a great deal of motivation to see OpenStack widely adopted, and they have zero incentive to deny contribution back to the community. This positions OpenStack to grow into a mature and well-used technology product.
There’s a Budget.
Rackspace is well on its way to having its own internal organization to support OpenStack, and NASA has its own independent budget for supporting the effort. Startups such as cloud.com have announced that they intend to support the product. CloudScaling, the company building many of the new clouds coming online, already has job postings available on the OpenStack community site. On top of all that, there are another 25 companies, including my own, which have a deep interest in contributing. Between Rackspace, NASA, Citrix, Dell and VC-fueled startups, there are already millions of dollars in capital being deployed toward its success.
Code is available today, and it’s the most important part of the OpenStack story. The project is built with the right foundation. The software is Apache 2.0-licensed, there are mailing lists, bug trackers, IRC channels (yes, those are still important in the free software community), and — most importantly — patches are already being accepted. Right now, a developer can pull down the source from Launchpad, read the wiki, and get the environment set up. If OpenStack has a shot at becoming a cloud computing standard, this is the foundation to make it happen.
But There’s Still Risk.
While OpenStack is off to a strong start, no initiative exists without risks. Overnight, Rackspace transformed itself from a hosting company to a development shop — and an open source development shop at that. Organizations such as the Apache Software Foundation and the Mozilla Foundation have spent decades solving the intricacies of open source community management. OpenStack is forming its own organization to deal with this aspect of the project, and building a strong and cohesive community is not an easy task. However, Rackspace has made some key hires — such as Rick Clark and Soren Hansen of Ubuntu Server fame — to help mitigate this risk. While there’s a good chance OpenStack will reach its goals, Rackspace should continue to invest in making sure the management team supporting the project has strong free software experience.
Serving too many business interests at once also poses a threat to OpenStack. Right out of the gate, the project is trying to solve NASA’s enterprise use case alongside Rackspace’s public service providers use case. These objectives are different enough that they could warrant independent projects. If managed incorrectly, OpenStack could end up unable to meet the requirements of either, leaving it crippled.
However, it’s hard to imagine a better scenario than that announced last month, so if OpenStack can’t fulfill the promise of an open cloud, then I’m not sure what will. The cloud itself was built on an open kernel and open hypervisors, so it’s only natural for a strong open cloud to emerge. With the success of OpenStack, we’ll have the standardization that’s been talked about for years. With it, we’ll be able to conclude our own initiative, Libcloud, and that’s the best sign of success we could have hoped for.
Related GigaOM Pro Research (sub required):
Alex Polvi is the CEO of Cloudkick, a San Francisco based cloud management company. Cloudkick is the force behind the Apache Incubator Libcloud, and provides tools that allow customers to easily build environments on many different cloud providers.