Open ... and Shut The cloud is the new operating system, and Amazon owns the cloud. Big iron vendors like IBM and HP are feeling the heat as workloads itch to move off expensive mainframes into Amazon's public cloud. Even Microsoft, the once undisputed king of the operating system, is under siege as its Seattle neighbor embraces and extends .NET and SQL Server, making Amazon, not Microsoft, the one-stop shop for enterprise computing needs.
Given Amazon's appetite and momentum, it remains to be seen which of its business partners will benefit from its rise, and which will be swept aside as it incorporates more and more functionality into Amazon Web Services.
Canonical, primary sponsor of the Ubuntu Linux operating system, may fall into both camps.
Amazon chief technology officer Werner Vogels notes that Amazon has had a profound effect on the startup community, offering robust infrastructure for a comparative pittance. Canonical, to a lesser degree, is much the same. Many of the startups I know use Ubuntu (on AWS) as their chosen cloud platform. Based on these data from a Cloud Market survey, it looks like they're not alone:
Indeed, while Red Hat dominates the data centre, every survey I've seen suggests that Ubuntu is still the dominant cloud OS, above RHEL, CentOS, and even Windows. This is the case for AWS, Rackspace, and other public cloud infrastructure. It's also true of the market for public websites where Ubuntu also surpasses Red Hat.
Red Hat cloud chief Scott Crenshaw, however, disputes the notion that Canonical's cloud lead is anything more than "fluff" based on cherry-picking "selective data":
When it comes to heavy lifting - building 'serious scale' clouds - spending hundreds of millions of dollars on proprietary software and making it open source, they [Canonical] are not anywhere in the same ballpark as Red Hat. When companies want an operating system platform they can rely on they come to Red Hat for a few reasons. The first is we've got first-class world wide support, the second is we actually write a lot of code unlike Canonical, so who better to fix or prevent problems... Of all the things that keep me awake at night, Canonical is not one of them.
Both Canonical and Red Hat may be right.
As the former chief operating officer at Canonical, I saw the company's revenue rise considerably, with a significant portion of that growth coming from growing enterprise adoption. But while I was there (2010), that uptake was coming from the data center, not the cloud. Cloud was still a nascent market: no one except Amazon was making any money there.
This may have changed, but my hunch is that the AWS crowd still treats Ubuntu like a free lunch - one that they're happy to "eat" in large quantities given its quality. HP's new public cloud, for example, is built in part on Ubuntu, and Ubuntu is a first-class guest OS citizen. But Canonical will likely get paid $0.00 for every Ubuntu guest OS deployed.
The question, therefore, is how to turn all of this adoption into cash.
Amazon makes money by charging for utility-like compute services. Red Hat piggybacks on this by charging for certified versions of its popular Red Hat Enterprise Linux software. Canonical, however, has a longstanding commitment to not package Ubuntu in this way, which is both a blessing to adoption and a curse to monetisation.
As Amazon continues to consume a rising number of enterprise workloads, Canonical needs to figure out a way to march in lockstep. This almost certainly will require a modification of its current licensing and/or pricing model. Amazon likes to run an amazingly low-margin business, and any costs that it can shed it will. Ubuntu makes a great guest OS because it's free, which means more money can be spent on AWS. Were Canonical to demand a cut of the action for all its Ubuntu adoption, I'm positive the response would be: "Next in line... CentOS?"
In other words, Canonical can't rely on the good grace of Amazon, or anyone else. It needs to find a way to insist upon payment.
Fortuitously, the cloud is ready-made for open-source monetisation. Software offered as a service is easy to charge for. The trick is figuring out which services startups and enterprises need to get the most from their cloud-based Ubuntu deployments. Juju seems like a good start. But more than the current crop of offerings is needed to really sell into the enterprise.
In sum, Canonical has nailed one hard problem: adoption. The second hard problem - monetisation - is looming [Sun Microsystems faced a similar challenge on Java - ed], but Amazon actually provides both a threat and an answer: services. If Canonical can crack the code on appropriate services to complement Ubuntu cloud deployments, it will have a massive business on its hands. ®
Matt Asay is senior vice president of business development at Nodeable, offering systems management for managing and analysing cloud-based data. He was formerly SVP of biz dev at HTML5 start-up Strobe and chief operating officer of Ubuntu commercial operation Canonical. With more than a decade spent in open source, Asay served as Alfresco's general manager for the Americas and vice president of business development, and he helped put Novell on its open source track. Asay is an emeritus board member of the Open Source Initiative (OSI). His column, Open...and Shut, appears three times a week on The Register.