OpenStack Summit 2014 In Atlanta, where the drinks are syrupy and the air hangs over you like a wet, warm blanket, four and a half thousand people have assembled to work on a crazy idea.
The idea? That a band of companies with different allegiances and areas of focus can come together to build some freely available tools that will let you, me, and everyone we know take on private companies such as Amazon, Google, and Microsoft.
These people are all crazy because the OpenStack technology they're working on is aiming to do to cloud computing what Linux did to operating systems – that is, provide a stable, sensible, free alternative to proprietary stuff.
This is an extremely difficult thing to do, and four years in, OpenStack is only now showing glimmers of the riches its founders NASA and Rackspace promised the community would find in the tech.
The cloud management, provisioning, and control tech has had these problems because where Linux only had to worry about the arcane intricacies of an operating system on a single chip with a few cores [Only, hah!—Ed.], OpenStack needs to marshal the resources of thousands of chips and tens of thousands of ports and servers across multiple data centers. This is a tough task, and one that companies like Google, Amazon, and Microsoft have spent over a decade building proprietary tools to tackle.
The OpenStack community believes in itself, though, and casts itself as an upstart guerrilla movement against the large incumbents.
"We have a lot to celebrate as the rebel alliance," said Troy Toman of Rackspace in a keynote speech on Monday. "We may not be facing the Death Star, but we're far from being able to declare victory on what we're trying to do here. We really just earned the right to compete, and we know some large proprietary cloud vendors won the right to too."
(Long-time members of the IT industry may remember that Sun Microsystems tried this "rebel" tactic against Microsoft, to little effect.)
"This is really a vision that is much bigger than one single cloud that battles it out feature by feature with everybody else," said Toman. "What we're really talking about here is the concept of cloud computing. I believe what we're really building here is the backbone for that architecture of the future."
That architecture of the future, if OpenStack has success, will be a single cloud operating system that can be run on small amounts of machines right up to systems dealing in thousands of servers and millions of connected devices.
So far, OpenStack has failed to deliver on these lofty goals, with few public cases of major usage. The telco Ericcson recently committed to a five-year, $30m OpenStack rollout with integrator Mirantis, and the Walt Disney Company came on stage on Monday to talk about how it used a managed OpenStack service from startup Metacloud in its own IT.
We are yet to see a huge Google-scale use case of the tech.
Wells Fargo, for example, came on stage on Monday to say that it had adopted the technology for some of its internal private cloud. The company had been moved to publicly say it was using the tech to help give a boost to the community, said Wells Fargo's head of private cloud enablement, Glenn Ferguson.
"This is my form of contribution, letting the community know this is very valuable, letting the community know what the use cases are, and that we're trying to run a serious business on this technology," he said. What Ferguson neglected to mention was how and where Wells Fargo is using the technology.
During lunch, some Red Hat employees told your correspondent that interest in OpenStack over the past year or so had climbed massively, but the company is still very far away from any major production deployments. This lines up with Red Hat's own on-the-record comments to us that it reckons it'll wring big cash from the tech in late 2015.
In other words, like nuclear fusion, a large number of paying OpenStack customers are just round the corner, same as it ever was.
The central issue facing the technology, based on speeches and snatched chats in corridors on Monday, is that it hasn't matured to the point that an enterprise is willing to bet everything on it, but the only way the technology can mature is if someone bets on it.
This was illustrated in a technical session on the system's hugely troubled "Neutron" networking component. While the panel talked about strategy and goals, the questions from the audience all related to the lack of usability in the system.
"When we talk to end users, their biggest pain point is in the networking," said Chris Wright of Red Hat. "It's not too lofty to say let's focus on core stability and core functionality."
Wright's comments contrasted with the early speech by Toman who had argued that the OpenStack community shouldn't focus on core features at the expense of further invention. "History has convinced me that's the wrong point of view," Toman said. "Linux didn't survive just because of Linux, but because of an ecosystem that included the LAMP Stack."
But, we would point out, Linux was reasonably solid by the time the LAMP stack came along, whereas OpenStack still has huge problems in the core. Though the OpenStack Foundation and others regularly insist OpenStack does not need a dictator like Linus Torvalds to succeed, the confused messaging your correspondent saw on day one of the OpenStack conference told us the opposite.
In a choice between guerrillas and the state, you need to be a brave, brave organization to trust in the poorly-equipped loose-knit group seeking to upturn the world. The benefit, though, is that if they win, so do you – hugely.
Right now, however, making that leap seems to require more faith than rationality. ®