The container space is starting to heat up. For a quick review on containers—what they are and when to use them—review my colleague’s article here. Docker is starting to seem increasingly isolated in this space, which it dominated with its early mover advantage before appearing to get sideswiped by the later entrants.
VMware stole a large amount of Docker’s thunder by releasing VIC (vSphere Integrated Containers—now at version 1.3) and Photon. Then, Docker lost the container orchestration and management space in a three-way war between Docker Swarm, Mesos, and Kubernetes, a little company that came out of Google. It effectively gave up and announced that Kubernetes will be seamlessly integrated into the latest version of Docker EE. This was a couple of months after Apache, which runs the Mesos project, also gave up the fight and added Kubernetes. Further, PaaS providers Pivotal and Cloud Foundry also announced support for Kubernetes in their respective container products. Why did Kubernetes win over Docker Swarm and Mesos? It was obviously not simplicity, as simple it is not. Docker Swam orchestrates Docker; Kubernetes, once correctly deployed, can support many more use cases.
The latest hammer and nail pointed in its direction is the acquisition of the CoreOS project and company by that giant of Open Source, Red Hat.
For the uninitiated, CoreOS is a software company that provides, amongst other things, Tetonic, an automated Kubernetes platform (yes, it makes Kubernetes deployments simpler), and Container Linux, a lightweight distribution optimized for running containers.
I mean, didn’t Red Hat already have a container platform called OpenShift? Yes, it did, but the truth of the story is that it is not that good. CoreOS will neatly round off its container story. Red Hat was missing the ability to elastically scale its containers with ease. Orchestration, and by this I mean Tetonic, will enable Red Hat’s container offering to scale up and down with ease.
But Red Hat used Kubernetes, so why does it need Tetonic?
As already alluded to, Kubernetes, although a very capable product, is difficult to deploy and does not truly address the complexities of managing stateful and stateless containers at scale. Tetonic is a proven enterprise-deployable Kubernetes solution. The overlay that Tetonic provides to the complexity that is Kubernetes flattens out the pain threshold. There is a reason why Kubernetes won the orchestration battles: it’s because it is easy to use. There is a reason CoreOS was bought by Red Hat: it’s because Tetonic flattens out the pain threshold in deploying Kubernetes.
This acquisition moves Red Hat to a position of power in the hybrid and modern application cloud deployments arena.
However, Red Hat will not just use Tetonic for container management. Open Stack is another application with a complexity problem. So too is cloud management and deployment. With multicloud becoming a real need, cross-platform orchestration and management becomes more and more important. Kubernetes has cross platform in buckets. However, traditional Kubernetes deployment on differing platforms was problematic due to each platform’s having a distinct and differing installation and configuration process.
This acquisition is not about Red Hat’s traditional competitors in the ’NIX/Linux arena (IBM, HP, and Oracle) but more about dealing with Docker’s eating its lunch. If it pulls off the integration of the product without losing too many of CoreOS’s current customers, which I feel it has covered with its statement on honouring current support contracts, the $250 million price tag will be small peanuts.