The OpenStack drama and the future of the cloud

For some time, I’ve been fascinated by OpenStack, an exciting open source project that’s evolving into a “cloud operating system” for the data center. It was a shock last week to see Citrix Systems, a charter member of the OpenStack consortium, make a loud exclamation of no confidence and pull out, opting instead to create its own open source project based on CloudStack — Citrix’s own cloud software with similar functionality.

I suppose a dustup like this was to be expected. OpenStack is highly ambitious. A collaborative project launched by Rackspace and NASA, OpenStack boasts a large and growing pile of cloud orchestration services — virtual machine management, object storage, machine image management — plus, in its latest Essex release, authentication and dashboard monitoring services. Part of the idea is that anyone can adopt OpenStack, which is available under an Apache 2 license, and create their own OpenStack flavors, much as various providers have developed their own Linux distributions based on the Linux kernel. Citrix seemed to be headed in that direction.

[ Read all about the fracas in OpenStack cloud gains version, loses Citrix. | What’s the private cloud and why would you possibly need one? Read the Private Cloud Deep Dive Report by InfoWorld’s Matt Prigge. ]

But no, says Citrix, we think certain components of OpenStack are too immature and evolving too slowly. We think a cloud operating system is a great idea, and so do our customers, but they want to get up and running now, not when OpenStack grows up. When I asked Citrix representatives whether management of the OpenStack project was part of the reason Citrix was bailing, I quickly got an affirmative answer, the clear implication being that OpenStack is a sprawling project with too many cooks and not enough central coordination.

Says Citrix: Jump on our CloudStack open source bandwagon instead.

Chris Kemp, CEO of Nebula and one of OpenStack’s NASA progenitors, sees things a little differently. Last week he offered InfoWorld contributing editor Oliver Rist this blunt assertion: “I think Citrix lied about aligning with OpenStack, then completely changed its position and viciously attacked OpenStack and threw it under the bus.” He believes Citrix was actually more invested in Cloud.com, which Citrix bought last year for $200 million and now provides much of the core code for CloudStack.

I am in no position to determine Citrix’s real motives or to say which has the better technology — at least until the InfoWorld Test Center does some gargantuan comparison test. On one level, you could say this imbroglio was really between Citrix and Rackspace, two second-tier players busting a move in the cloud and colliding. Kemp claims “there’s no one outside of Citrix contributing to” CloudStack in the open source community. But Rackspace admitted to me that “about 60 percent” of the code for OpenStack was contributed by Rackspace itself, including, presumably, the code it acquired from NASA.

Nonetheless, both are now open source projects — and the fact that open source is blazing a trail in such an important new area is pretty stunning when you think about it. But it’s not happening in a vacuum. EMC VMware is assembling its own complicated commercial cloud solution, which extends all the way to application management; given the company’s long lead in virtualization management, it may well arrive at a fully baked solution first. Microsoft is coming on strong with its own armada of private cloud software.

For some enterprises, the messiness of last week’s split will reinforce fears of legal land mines in open source code, even though the two have different code bases. Those customers would probably be inclined to pay the licensing fees and go with the commercial cloud solution they’re most comfortable with anyhow. Others may be drawn to Citrix or Eucalyptus or whoever emerges as the OpenStack equivalent of Red Hat. But I think OpenStack’s notion of an open source cloud operating system as ubiquitous for enterprise private clouds as Linux has become on enterprise servers remains compelling — and will be even more so when, one day, the technology to manage hybrid private/public clouds as a continuous fabric finally emerges.

You can bet that most public cloud service providers will take the open source route like they always have. Years from now they can say to customers: Adopt the same cloud orchestration software on premise and you can seamlessly burst to us. That’s not even close to being true today, but one day it will be.

Sure, the big incumbents like Amazon.com, Google, and Salesforce.com have developed their own cloud orchestration — but many other providers will opt for OpenStack, CloudStack, or some other open source solution. When you really have to scale, how much of your profits do you want to go to VMware?

Yes, Citrix did make a small dent in OpenStack. But OpenStack has too much momentum to grind to a halt. Who knows which cloud orchestration scheme will ultimately dominate, but I’m pretty sure it will be open source in the end — if only because the march to the open source public cloud is inevitable.

This article, “The OpenStack drama and the future of the cloud,” originally appeared at InfoWorld.com. Read more of Eric Knorr’s Modernizing IT blog, and for the latest business technology news, follow InfoWorld on Twitter.