Open Source

Red Hat can't afford to let HP lead OpenStack

Red Hat understands the world is moving to mobile and cloud, which is why it can't allow HP to claim the OpenStack top contributor distinction.

Cloud security

Red Hat owns the enterprise Linux market, but that's yesterday's news and yesterday's market. As Red Hat CEO Jim Whitehurst recently declared, "We're in the midst of a major shift from client-server to cloud-mobile." The problem for the Linux giant is that it still lacks a dominant presence in enterprise cloud and has hardly any footprint in mobile.

In other words, Red Hat has a lot of work to do.

Getting the "vision thing" right

Importantly, Red Hat recognizes what it's up against. This week, Whitehurst clearly identified the opportunity and also the challenge facing the enterprise open source leader:

"We're in the midst of a major shift from client-server to cloud-mobile. It's a once-every-20-years kind of change. As history has shown us, in the early days of those changes, winners emerge that set the standards for that era — think Wintel in the client-server arena. We're staring at a huge opportunity — the chance to become the leader in enterprise cloud, much like we are the leader in enterprise open source. The competition is fierce, and companies will have several choices for their cloud needs. But the prize is the chance to establish open source as the default choice of this next era and to position Red Hat as the provider of choice for enterprises' entire cloud infrastructure."

So, the company acquired Feedhenry, a "backend platform that enables the mobile user to access data, build backend logic, and access corporate APIs, all in a scalable, secure manner." It's a smart move for a company that needs to compete in mobile but is never going to be the foundation for the next Angry Birds.

Of more concern is the area where Red Hat should be strong and ascendant: cloud.

According to TheCloudMarket.com, which measures operating system adoption on Amazon Web Services (AWS), Canonical's Ubuntu leads with 54% of the market. Of more concern, however, is that 55% of OpenStack users also prefer Ubuntu, according to a survey of OpenStack users. OpenStack is Red Hat's home turf. It can't afford to lose at home.

Making OpenStack work

Mobile is a hard market and arguably hinges on Red Hat first claiming the cloud. Yet Red Hat, along with virtually every incumbent enterprise vendor, has been slow to shift to the cloud even as AWS roars into the enviable position of being five times as big as the next 14 biggest cloud providers combined as of 2013. While the AWS lead has almost certainly narrowed a bit with the rise of Google and Microsoft, it's still formidable.

While Red Hat is unlikely to make a dent in AWS' dramatic lead in public cloud computing, it doesn't really need to. Red Hat's strategy is to ensure that it provides a consistent platform that spans every major public cloud and ties these public resources to OpenStack-based private cloud resources behind the firewall.

Red Hat's strategy is an "and" strategy, not an "or" strategy.

And yet Red Hat, once the clear leader in OpenStack development, has fallen behind to HP, as Stackalytics contributor data for the OpenStack Juno release shows:

Red Hat has fallen below HP

This is a big change from the Icehouse release, when Red Hat accounted for 19% of all contributions and HP generated just 13%, or the Havana release when Red Hat had 17% of all contributions and HP offered 14%.

While one release isn't dispositive, it's also not a trend Red Hat can afford to start.

Code, after all, is currency in open source. The more code you contribute, the more you influence or control a project.

Let's be clear: OpenStack needs a leader, and I've long argued that OpenStack has a history of undisciplined product governance. The best open-source projects, like Linux, have a strong, small leadership team that knows the value of saying "No." But OpenStack has struggled with the "no" word, thereby delivering a hodgepodge of features of varying quality and doubtful integration.

That's part of the problem with a community: you get all sorts of definitions of what the product is, and that can lead to disorder in the absence of strong leadership, as Gartner has pointed out.

The long game

Red Hat isn't doomed. Not even close. As I mentioned, one release does not a market make, especially in a market that is still in its infancy.

But Red Hat needs to keep hold of its best people (the company recently lost its CTO to Google) and add to its army. HP recently acquired Eucalyptus and, in the bargain, got its exceptional CEO, Marten Mickos, who now runs HP's cloud business. Give Mickos a few quarters of code dominance of OpenStack contributions, and he'll find a way to turn that into revenue dominance, too, as he did with MySQL.

The great news in all this is that, as companies like Red Hat and HP fight to contribute the most and best code, enterprises win.

The other great news is that Red Hat is increasingly focused on the developers that will drive its adoption in the cloud and mobile, whether through increased OpenShift investments or acquisitions like Feedhenry. Red Hat, in other words, knows what it needs to do. The key now is execution.

About Matt Asay

Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.

Editor's Picks

Free Newsletters, In your Inbox