OpenStack Success Disputed as Backers Challenge Technical Direction
Last week marked the third anniversary of the OpenStack project, an effort led by Rackspace and NASA to create an open source cloud operating environment. OpenStack quickly gained momentum and has evolved as a huge force in cloud computing, with 235 member companies that include AT&T, IBM, Red Hat, Cisco, Hewlett-Packard, Rightscale, Internap and Mirantis.
Attendance at the semi-annual OpenStack Summit continues to increase exponentially, and the OpenStack Foundation claims enterprise adoption is growing, citing examples such as PayPal, Cisco WebEx, Best Buy, Bloomberg, the Gap and HubSpot, as well as the recently reported deployment by Fidelity Investments.
However, some prominent critics have questioned whether OpenStack is gaining meaningful adoption compared with the growth of Amazon Web Services, Google Compute Engine and Microsoft's Windows Azure, among others. Analyst David Linthicum noted in a GigaOM blog post that despite a strong ecosystem and buzz for OpenStack, overall adoption pales in comparison to the growth of AWS.
"While OpenStack, including Rackspace, HP, IBM, and many startups, is clearly the darling of the cloud tech community, the number of installations within traditional IT shops has been lackluster when you consider the expectations that were set," Linthicum wrote.
Perhaps the most noteworthy critic to take the wind out of OpenStack's sails was Randy Bias, the outspoken CEO of Cloudscaling, itself a founding OpenStack member. Bias posted an extensive critique of the existing technical agenda outlining why he believes the OpenStack Foundation's self-described native APIs lack true compatibility with Amazon Web Services APIs, which was the project's original mission. This is especially important now considering the widespread use and dominance of Amazon's cloud services. The OpenStack Nova compute APIs are largely identical with the Rackspace Cloud Servers public cloud service API, not Amazon's, Bias said.
"There is nothing 'native' about the Nova API," Bias wrote. "In fact, calling the Rackspace Cloud Servers API the 'native API' promulgates the notion that there is an OpenStack Nova API that is separate from Amazon's. It's now obvious that the original native API for OpenStack was in fact its AWS EC2 API."
Now that Rackspace has ceded control of the project to the OpenStack Foundation, the new governing board needs to revisit the API stack, according to Bias. "In short, the community controls the direction of the project, and it's time we advocate a public cloud compatibility strategy that is in all our best interests, not just those of a single, albeit substantial, contributor," he wrote. "Failing to make this change in strategy could ultimately lead to the project's irrelevance and death."
The reason, he contends, is that Amazon is far more dominant than any other public cloud Infrastructure as a Service (IaaS). "Embracing Amazon serves the interests of all community members by positioning OpenStack as the best choice for enterprises and SaaS providers that want an ecosystem approach to public cloud, one in which their applications can move to the infrastructure best suited to the job at that time," he said.
Also, despite the lack publicly disclosed information, Bias believes the recently released Google Compute engine IaaS is also growing rapidly. "If others arise, we should debate and evaluate embracing them only when their market position is established," he argued.
Specifically, Bias proposed the OpenStack Foundation do the following:
"1. Embrace major public cloud APIs. GCE, AWS, Azure, and possibly vCloud
"2. Rename the Nova API to the Rackspace Cloud Servers API
"3. Create a new low level API(4) and move to the bridged API model
"4. Expand testing and the work around refstack. Refstack should focus on public cloud interoperability & hybrid cloud
"5. Embrace existing AWS interoperability testing frameworks."
I reached out to officials at the OpenStack Foundation and Rackspace, and while I didn't hear back at press time, I did speak with IBM distinguished engineer and CTO for cloud interoperability Chris Ferris today, primarily to discuss Big Blue's decision to commit to the Cloud Foundry Platform as a Service (PaaS) stack, originated by VMware and spun off to the new Pivotal. Regarding Bias' post, Ferris said he disagreed.
"I've known Randy for a while and he's a very bright guy," Ferris said. "I respect his opinion but I disagree with his conclusion. It's not at all clear to me Amazon has won anything, but more importantly, IBM really believes firmly that open is the right way. Adopting a proprietary API that is in the exclusive control of one vendor is not open. If Amazon wants to contribute that and make it part of OpenStack under the Apache 2 license, maybe we would think about that. But the notion that we should cede the whole thing to Amazon is not my idea of a good idea."
Posted by Jeffrey Schwartz on 07/25/2013 at 12:49 PM