Sorry, Linux. Kubernetes is now the OS that matters

Linux is just plumbing. The real OS—the real value—is with Kubernetes

Sorry, Linux. Kubernetes is now the OS that matters
Getty Images

The operating system no longer really matters. And for developers and the cloud, that means that Linux no longer really matters.

You can see proof of that in what has not happened. Canonical, the company behind Ubuntu, has not gotten a $34 billion buyout offer from IBM, even though company founder Mark Shutteworth would have taken that deal, despite protestations that the company isn’t looking for a buyer.

The reason that a deal wasn’t offered to Canonical, and won’t be anytime soon, is because the technology industry no longer values the operating system. Or, rather, the tech industry has a new operating system it values: It’s called Kubernetes.

We now live in a Kubernetes world

Perhaps Redmonk analyst Stephen O’Grady said it best: “If there was any question in the wake of IBM’s $34 billion acquisition of Red Hat and its Kubernetes-based OpenShift offering that it’s Kubernetes’s world and we’re all just living in it, those [questions] should be over.” There has been nearly $60 billion in open source M&A in 2018, but most of it revolves around Kubernetes. Red Hat, for its part, has long been (rightly) labeled the enterprise Linux standard, but IBM didn’t pay for Red Hat Enterprise Linux. Not really.

No, what IBM spending an incredible premium to get was a Kubernetes-fueled cloud clue.

Over the past several years, Red Hat’s Kubernetes-based OpenShift offering has driven continued revenue growth. Yes, RHEL accounts for 64 percent of Red Hat’s revenue, but it’s growing at a mere 8 percent while OpenShift is powering forward at multiples of that growth rate.

Moreover, OpenShift deals actually protect and extend RHEL revenue by including RHEL in the subscription. This happens naturally because, as Red Hat CEO Jim Whitehurst acknowledged on the last Red Hat earnings call, “big data workloads run on Linux. AI workloads run on Linux. Devops and those platforms almost exclusively Linux. So much of the net new workloads that are being built have an affinity for Linux.”

But, again, it’s not Linux that is interesting. It’s just plumbing. The real OS—the real value—is with Kubernetes.

Kubernetes is the new enterprise Linux

Red Hat’s Daniel Riek stated this most forcefully last year, declaring that Kubernetes is the new enterprise Linux:

AWS, Microsoft Azure, Google Cloud are the new EMC, HP, Cisco, Sun, and Oracle. And like RHEL provided an open alternative to the vertically integrated, proprietary mainframe and Unix systems, and allowed customers to choose from an open ecosystem and hybrid infrastructure, [Kubernetes] offers an alternative to the vertically integrated proprietary cloud. In this context, [Kubernetes] is the new [operating system].

In this world, it really doesn’t matter (at all) that Ubuntu is the most popular OS instance running on Amazon Web Services, as Cloud Market measures. If it did, it would also matter that Canonical is responsible for orders of magnitude more OS images than AWS. But of course that’s a ridiculous comparison: AWS mints nearly the price of Red Hat in revenue each year, while Canonical is still measuring revenue in tens of millions. For AWS, there is comparatively little revenue in supporting Linux images. Rather, the money comes from virtualization and container-driven services that yield higher value to developers.

As an industry, we’ve decided to build on the next operating system, and it’s called Kubernetes. There’s still money to be made supporting old-school Linux or even old-school clouds (built with OpenStack). But that’s a backward-looking way to scrape together pennies when Kubernetes is going to be worth hundreds of billions in value, or more.

Kubernetes will one day be a faded OS too

Sure, eventually Kubernetes, too, will fade in visible importance, as Kubernetes cocreator Brendan Burns recently opined:

I view in the Kubernetes and the Kubernetes APIs as being like Posix. Every program that you ever run on a Linux system runs via Posix APIs, but you don't really think about them very much. You learn them in operating systems, and maybe you do some Pthreads or whatever, but you don't think about them very much. I hope that that's where we end up with Kubernetes as well; that it fades to the background. It's important, and it's useful and it's something that is backbone of everything that you do, but you're thinking about higher level abstractions.

In other words, Kubernetes will become like an operating system—and like the Linux that used to be the center of attention.