The expansion of OpenStack to include a wider range of development projects – known as the ‘big tent’ approach – will fall by the wayside as it becomes the platform of choice for private clouds, says Canonical founder, Mark Shuttleworth.
“Most of these bullshit as a service big-tent things are going to die and go away,” he said. “What will remain is core infrastructure as a service. It is extremely important for every CIO. It is a foundational piece on which private cloud and public cloud will be essentially balanced against one another.”
Shuttleworth likened the growth of vendor interest in OpenStack to the dotcom boom and bust, with the open source platform emerging leaner and more focused.
“It’s a bit like the dot com bomb,” Shuttleworth said. “'The internet will fail!' No, it didn’t. A whole bunch of stories failed, mythology failed, but the internet kept going. It’s become the de facto standard. Very similarly, a lot of the mythology around the ‘big tent’ of OpenStack will go bust.
“All the things people say next year will be great – will die. That’s healthy. That’s a cleaning out of focus on what matters.”
Canonical’s OpenStack platform on Ubuntu is winning the company business, and a re-focusing of clarifying OpenStack’s core proposition could well win it more.
He insists that the heart of OpenStack is “very important” and is “doing very well”.
“We’re able to show a very quick return on investment to large organisations looking for internal infrastructure as a service," said Shuttleworth. "And we're also able to show quick engagement with their developers, and productivity gains effectively for their development teams, and show comparable economics.”
“The real question is can people stand up an internal OpenStack cost-effectively.”
He also floats the idea that there is perhaps more to the elevator pitch of avoiding vendor lock-in than the vendors themselves are admitting.
“Vendors will play a game where they say: ‘oh, you’re not locked in, it’s all open’ - but the truth is, if your developers are writing to an API you can only get internally, you’re locked in to your internal IT. No one wants that.”
Shuttleworth believes the reality of OpenStack adoption moving forward will be driven by multiple clouds, before adding: “Go ask a CIO.”
What next for OpenStack, then, if the “circus” tent – as Shuttleworth puts it – is due for a collapse? He believes there will be a confidence crisis and a blame game.
When asked about the importance of integration between OpenStack and AWS, he said that “part of the collapse of this bubble of bullshit will be a blame game saying OpenStack failed”.
“[People will say] 'OpenStack did the wrong thing, OpenStack made the wrong choices, OpenStack should have copied the AWS APIs',” he said. “That’s all these guys saying ‘we’re not getting what we thought we would get so it’s his fault'. It’s not.”
“Part of that blame game will be an existential crisis over APIs but it doesn’t really matter. If you put yourself in the mind of a CIO you must commit – they are committing to working across multiple clouds. Having AWS and Azure makes sense: once you do that, having OpenStack as well is not hard.”
“I think the API ship has sailed, there were smarter decisions that could have been made five years ago, we’re all human – it doesn’t matter.”