Many enterprise software companies try to jazz up their wares with fancy buzzwords (Cloud! Machine learning! Augmented reality!). Not Red Hat. Red Hat plays a very different game. Instead of trying to make software exciting, its stated goal is to make software very, very boring, as Red Hat Chief ARM Architect Jon Masters told a crowd at Supercomputing 2016.
By “boring” Masters didn’t mean uninteresting. Instead, as he went on to explain, Red Hat “make[s] stuff that just works.” This message plays well to risk-averse CIOs.
It’s not hard to see how this mantra applies to Red Hat’s booming businesses in operating systems, middleware, and storage. But Kubernetes? Yes, that’s right: Red Hat aims to bring the same boring “just works” story to the still-nascent world of containers which, ironically, may be the most exciting thing to ever happen to the hype-heavy world of cloud-native applications.
Let’s not call it ‘boring’
Red Hat doesn’t have a history of embracing the novel. The company commercialized Linux when it was still relatively early (though the operating system market was quite mature), yet didn’t jump into other areas like middleware (JBoss) until those markets were ripe for disruption. As such, Red Hat’s earnest efforts to dive into Kubernetes signal a changing strategy. (For those who would point to its OpenStack adoption as another such signal, it isn’t. OpenStack is best read as a way to convince CIOs that they’re embracing the cloud, while keeping them firmly grounded in their comfortable data centers.)
So, Kubernetes is different. Kubernetes bills itself as “an open-source platform for automating deployment, scaling, and operations of application containers across clusters of hosts, providing container-centric infrastructure,” which certainly sounds dull and, hence, perfect for Red Hat. But, given that enterprise adoption of containers is still in its infancy, the very mention of the word “containers” makes Kubernetes pretty avant-garde by Red Hat standards.
Not that Red Hat is calling Kubernetes “boring.” Instead, they’re calling it “Enterprise-Ready,” which is basically the same thing.
Indeed, Red Hat CTO Chris Wright talked up the company’s PaaS offering, OpenShift, as “enterprise-ready Kubernetes” in a keynote address at KubeCon 2016. Kubernetes is far and away the most popular container management tool but, as Red Hat’s Brian Gracely captured, “companies [are] in a transition from a series of DIY Kubernetes projects to a new reality around managing them as a platform that would support production applications.”
The market, in short, is ripe for “boring.”
To be clear, Red Hat’s OpenShift, which is built on Kubernetes, is still an emerging growth story for Red Hat. As the company noted on its most recent earnings call, “large sophisticated customers are really gravitating towards OpenShift,” with particular traction in financial services. This jibes with Docker CEO Ben Golub’s comment to me earlier this year that “there is ample evidence to show that [containers are] being embraced by early majority/pragmatist organizations.”
What Red Hat is doing is lowering the bar to container adoption by making Kubernetes…boring.
As the no. 2 contributor to Kubernetes (behind Google), and the #1 enterprise software contributor, Red Hat has put itself in a good position to guide the future of Kubernetes, making it ever more enterprise-ready.
This strikes me as a much more appealing, if longer term, approach to winning the virtualization market. Yes, containers are arguably more complement than competitor to virtualization, and Red Hat offers both. But, long term, containers have the potential to up-end the traditional virtualization market, and Red Hat’s early foray into Kubernetes via OpenShift gives it a chance to shake up the industry in a way that Red Hat Virtualization simply doesn’t.
That is, provided Red Hat can make Kubernetes boring.