OpenStack to Be Production-Ready by JanuaryOctober 21, 2010
Calling upon founding member Rackspace’s Texas roots, the OpenStack project has released the first full version of its open source cloud-computing platform, dubbing it the “Austin” release. The new release includes a handful of improvements to the storage-platform code, OpenStack Object Storage, as well as the debut of OpenStack Compute – the Rackspace-Cloud-Servers-meets-NASA-Nebula compute engine.
According to OpenStack GM Jim Curry, the storage platform, which is based on Rackspace’s Cloud Files code, has been production-ready since the beginning, while the compute platform is currently ready primarily for testing and development purposes. However, a planned January release (code-named “Bexar”) will be production-ready for smaller deployments, and, in April, Curry expects OpenStack to release a version ready for large-scale production environments within corporate or service-provider data centers.
As Om explained upon the OpenStack launch in July, the aim of the project is to give service providers, individual organizations and technology vendors the tools they need to build their own clouds without exposing themselves to the costs and lock-in potential of buying proprietary software or writing their own code. It’s actually the burden of writing one’s own software that led Rackspace to go down the OpenStack road in the first place.
Curry says that although Rackspace traditionally made its money by using the same software as everybody else and adding service on top of it, that strategy didn’t work out with cloud computing. So, after the company wrote its own code to get Cloud Servers and Cloud Files up and running, it decided to reassess the situation. “We made a decision,” he explained. “‘Do we really want to continue [acting like a] software company with us being the sole consumer of that software, or do we want to figure out a way to get the world involved in this and have everybody contribute to it and make it a much better product, and build our hosting business on top of that?’
Curry gave a frank response when asked about the potential conflicts of interest inherent in contributing to an open cloud-platform project while also selling its own service – something that Rackspace and several contributors, such as Cloud.com, do. All Rackspace wants, he said, is access to the code, because Rackspace will always make its money by adding value to best-of-breed software. As for contributing vendors, they’re encouraged to at least give back to the project, but there’s nothing in the Apache 2.0 license that requires them to do so. He added that most community members have been very helpful thus far, with Citrix, in particular, adding much code to the Austin release and building OpenStack interoperability into its own products.
If OpenStack catches on like Curry thinks it will, it will be in members’ best interests to have tight ties into the OpenStack code. Compared with proprietary cloud platforms that have limited scope and interoperability, an ecosystem of OpenStack-powered public clouds, private clouds and software solutions could be a compelling proposition. The project is even courting enterprise-user contributors to ensure the product lives up to their expectations. As I’ve noted in the past, however, attaining such broad support won’t be easy given the growing number of competitive offerings.
Still, Curry is optimistic, in part because he says his phone has been ringing off the hook since the project launched. It’s a lot easier to get organizations active in testing out OpenStack when he’s “not trying to pitch a business or sell it to people.”