The OpenStack gold rush

Off the Beat: Bruce Byfield's Blog

Working with free software is like living with a mad carpenter -- each time you look around, you discover a room you knew nothing about. After four days at the OpenStack Summit in Vancouver, I feel like I've discovered an entirely new subdivision. And the best news is that OpenStack is reasserting the core philosophies of free software, although from a practical perspective rather than an idealistic one.

Like many people, I had been vaguely aware of OpenStack for several years. Something about cloud computing, wasn't it, and maybe containers? But until I attended the conference this week, I was unaware of what a major project OpenStack had become. According to the figures I heard, the Summit attracted over 6,000 attendees at $900 per badge, taking over all three and a half floors of the Vancouver Convention Center for five days with as many as fifteen programming tracks during peak periods. That is a scale that many LinuxCons couldn't match.

Another metric is the gold rush mood of the Summit -- the sense that something big is happening, and that now is a time to grab a piece of it. Of the 108 companies with a booth at the Summit, I counted over 80 that were less than three years old.

The rest were established companies trying to reinvent themselves with varying degrees of success, such as Cisco, IBM, Red Hat, SUSE, and Ubuntu. A gold rush-like sense of pride and hustle could be read in many of the slogans, such as VMWare's "The best platform for OpenStack, or in Ubuntu's offer to "upgrade" the Red Hat branded lanyards given out with the registration for Ubuntu branded ones.

Angel Diaz, IBM Vice President, Cloud, tells me that IBM made seven billion dollars last year -- and that was just one company. Cloud services and storage might be unglamorous compared to kernel or desktop development, but the money to be made from them must be staggering, and it was reflected by the lavishness with which that same money was spent at the Summit, with major tourist venues being taken over by Ubuntu and HP for evening events. All in all, it was an atmosphere I haven't sniffed since the Dot-Com era of 1999-2001.

A new face for free software
But what surprised me most was the philosophy driving the atmosphere. In the early days, the cloud was seen as a major challenge to free software, due mainly to the centralized social media sites like Facebook and Twitter. Despite the Affero General Public License, the Free Software Foundation was widely criticized for being slow to respond to that challenge. Proprietary free services, the conventional wisdom insisted, would mean the end of free software. I wrote as much myself.

When I mentioned that reaction during the Summit, all my listeners laughed -- and with good reason. OpenStack is becoming the major player in the cloud, and although case-studies discussed at the keynotes included Walmart and TD Bank, the spirit in which such successes were constructed might almost have come out of the mouth of Richard Stallman himself.

The difference, of course, is that those involved in OpenStack are grounded in practicality.
From the first, OpenStack has been an alliance of interests, so to work it has needed to encourage co-operation. Building on the example of the rise of Linux, it has gravitated towards an open source model because it is one that tech companies now understand, and because it works.

The result is that, while OpenStack is dominated by perhaps half a dozen companies, and some are obviously more committed to the project than other, the possibility of a centralized service like Facebook was never much of a possible model.
 In the latest development cycle, OpenStack is even talking about federation in Defcore, a common core of functionality that will enable all OpenStack deployments to communicate with each other and, with any luck, to produce a new level in interoperability. Both the concept of federation and its name are borrowed from Free Software Foundation-related projects like MediaGoblin, who consider federation as an alternative to the centralized social media sites. But where projects like MediaGoblin are trying to build federation of separate nodes, OpenStack's goal is to improve communication and functionality between already existing nodes. Where free software advocates have been dreaming, vastly out-gunned by Facebook and Twitter, OpenStack has been acting, and has the clout to enforce its approach. Suddenly, the example of the major threat to free software has emerged as its major ally, entirely for reasons of its own.
 

After the gold rush
  Gold rushes are followed by ghost towns, and possibly OpenStack will peak and even decline a few years from now, replaced by the next Big Thing. But then again, maybe not. While many of those eighty new companies I observed are likely doom to flame out, the needs that OpenStack is meeting and creating are unlikely to disappear. For all the echoes of the Dot-Com era, this is a gold rush with fewer airy dreams than its predecessors.
 

 Meanwhile, I keep remembering Angel Diaz's account of talking about IBM becoming involved with OpenStack with CEO Sam Palmisano. "Make it bigger than Linux," Diaz reports Palmisano as saying.  
 After exploring the world of OpenStack for four days, I wonder if that has already happened.