I don’t believe in Kubernetes
I have always been a believer in Cloud Computing. Even when big executives of the world were calling it a fad, I believed in the cloud—or its promise. It was not so much an achievement to be had—rather as with to so many other people in tech—it spoke to me from a place of understanding. We may not have called it “cloud”, but when Amazon opened the gates of their Web Services to everyone, we saw good things we recognized with and we realized it was an opportunity to build things better.
Cloud Computing was about getting good IT infrastructure accessible to the rest of us. Finally, we did not need to be a sizeable enterprise or have a small army of ops and devs just to have the necessary infrastructure in place. It was empowering, it was enabling. When cloud users could solve their problems conveniently, people interested in cloud infrastructure could enjoy building them—and in the process, change many professions and how the industry worked.
After achieving the cloud, our mission should have been to make cloud computing accessible to more and more people. We dreamt of people having on-demand access to all the IT resources they want at an ever reducing cost, being able to use technology easily and conveniently to solve problems they needed solving. That, should have been our mission.
Instead, here we are today left fixated on things like Kubernetes1 and trying to emulate how giant companies—so called hyperscalers—like Google and Facebook operate their infrastructure. Instead of trying to get people closer to solving their problems, we are effectively erecting newer barriers in complexity.
To be clear, I am not trying to diss Kubernetes. I like Kubernetes—with certain caveats. It is a system with many components and—as with any complex system—comes with inherent operational complexities. It is an interesting system catering specific needs. It has its place. More accurately, Kubernetes is a tool for cloud builders, not cloud consumers.
A problem with the current Cloud Native narrative is that a lot of companies (and in turn, a significant amount of eco system resources), who would otherwise be working with a consumer focus are busy with building tools for cloud builders—and then try to wrestle these into submission as cloud consumer technologies. You do not have to take my word for it—you can try deploying your next application to production on Kubernetes.
What happened to bringing the best technology conveniences to the rest of us? Whatever happened to harnessing the power of IT so that it stops being an inhibitor and becomes the great enabling force that it can be? What happened to—may Cthulhu have mercy on my soul—powering “digital transformation”?
When the most activity in tech is taking place on a tangent, it is refreshing to see ventures like Oxide Computer Company daring to pause and ask fundamental questions again. In their case—at least from what I can gather at this point—they are tying to bring the advantages of hyperscale IT infrastructure because the hardware industry seem to have forgotten the rest of us in favor of the hyperscalers.
As for me, I want modern, secure, cost-effective computing to be delivered to more and more people because our industry seem to have forgotten that along with the fixation on hyperscale. I do of course believe in the existence of Kubernetes. I just do not believe it is the thing that carries the mission forward—at least not in the current form. This definitely is something I would like to work on. It had always been my interest. At different times, in different places, in different forms—it had always been the mission.
PSA: Containers ≠ Docker ≠ Kubernetes ↩