Kubernetes is making waves in the container ecosystem and is changing how container vendors take their solutions to market. The most recent case that reflects this is Red Hat’s acquisition of CoreOS for $250 million. It’s the first in what could be a series of acquisitions and mergers toward more consolidation in the Kubernetes space. Let’s look at the drivers and motivations for this trend of consolidation and what it might look like in the coming months.
Kubernetes has its own challenges
Kelsey Hightower, a Kubernetes guru, wrote a highly useful guide titled "Kubernetes the hard way" a couple of years ago. However, at the time of that writing, and in many ways even now, there isn’t an alternative easy way to manage Kubernetes. Kubernetes brings complexities and unique challenges to container management. Yes, it abstracts a lot of low-lying issues with containers and provides a management layer that’s better than managing the underlying containers directly. It provides automation and intelligent defaults for various aspects like container networking, load balancing, service discovery, and high availability. It also has wide-ranging integrations for everything from cloud infrastructure management, to cloud storage options like S3 and EFS, monitoring, logging, and security. However, there are many aspects to consider when managing Kubernetes, and it takes a learning curve and getting familiar with the internals of Kubernetes. That’s the reason for the popularity of a guide like Hightower’s.
Explosion in Kubernetes service providers
What this shows is that there is an inherent complexity that characterizes Kubernetes and taming this complexity, and making it easier to manage, is the challenge of the hour with containers. This has led to a lot of activity to bring Kubernetes to the market, particularly to make it well-suited for enterprise operations. The last two years saw the rise of many new startups to provide managed Kubernetes services. This includes companies like Heptio — started by Kubernetes co-founders — Platform9, Apprenda Kismatic, Mirantis, and many more.
Some of these vendors are brand-new startups, but most are startups that have changed course and switched from OpenStack to focus on Kubernetes. The cloud vendors like Azure and AWS have also changed their tune, promoting container services from a second-class to a first-class citizen of their platform. They each rebranded their container solution to replace containers with Kubernetes. Azure Container Service (ACS) became AKS, and ECS became EKS. (The “K” is for Kubernetes.) This trend was started by Google, which rebranded from GCE to GKE. In fact, today it’s hard to stumble across a cloud vendor’s website without noticing that they support or are exclusively focused on Kubernetes as a Service.
All this activity has led to fragmentation in the market and a cluttered space of vendors duplicating effort with little to differentiate the leaders from the laggards. It has become confusing to pick the best Kubernetes solution out there. A survey by the CNCF shows that organizations running Kubernetes in production aren’t fully committed to just one Kubernetes service provider. In fact, many of these organizations use more than one solution. While Kubernetes on-premises is the most common mode of deployment, when it comes to cloud vendors, there’s little that separates one from the other in terms of market share.
This presents a situation where customers can easily jump ship from one vendor to another and not lose step. This is not a bad thing as it shows that there’s no lock-in. However, from the vendor’s perspective, this is a delicate spot to be in, and an opportunity waiting to be cashed in on. Vendors recognize this paradox and are desperate to create differentiation for their solutions. They can either build this Kubernetes expertise in-house or take the shortcut of acquiring and merging with like-minded organizations to form a complete solution. This is the backdrop for the CoreOS acquisition.
Finding common threads
In their press release, CoreOS CEO Alex Polvi talks about how Red Hat and CoreOS have been close partners since the early days of containers, working on many open source projects toward common ends. Evidently, CoreOS is a more prominent name in the open source ecosystem than Red Hat with its numerous valuable contributions. Most notably, CoreOS is closely involved in Kubernetes development, even heading the first non-Google-led release of Kubernetes early in 2017. It is among the top contributing organizations to Kubernetes development.
CoreOS rkt made a splash a few years ago when it was pronounced as a more open and secure alternative to the Docker runtime. It hasn’t gained adoption since then, but has been a key contributor to the Open Container Initiative (OCI) that promises to be a worthy alternative to the Docker runtime.
etcd is a distributed key-value database that’s a key component of Kubernetes and was created by CoreOS. etcd handles cluster operations within Kubernetes including vital underlying tasks like leader election between nodes. These kinds of operations were not possible or were performed manually before, but CoreOS shows its ingenuity in architecting systems for automation and smart defaults that don’t require human intervention and can operate with much higher efficiency.
CoreOS Container Linux is a minimalist container operating system that I covered in one of my articles a while back. It has automatic OS updates and a number of features that make container management easier. Container Linux formed the base for CoreOS’ main commercial solution — Tectonic. As a provider of managed Kubernetes, Tectonic is among the more popular options building on CoreOS’ close ties to the Kubernetes team at Google. CoreOS touts it as the enterprise solution for managing Kubernetes.
Apart from Tectonic, CoreOS also has Quay, which is one of the leading private container registries, and a great alternative to the public Docker Hub. It includes container image scanning and features to enable collaboration within an organization.
Throwing their hat in
Looking at this massive body of work, the motives for Red Hat’s acquisition of CoreOS become clear. First, CoreOS’ take on automating routine container management tasks is a big draw for Red Hat, which is looking to be the leading Kubernetes as a Service provider. Second, the numerous open source projects of CoreOS bring trust and reputation in a space where innovation is the most valued commodity. These open source projects come from organizations with a creative DNA that a more service-oriented company like Red Hat could use an infusion of.
Finally, the biggest reason for the acquisition is to leverage the talent and culture of CoreOS and infuse that energy into Red Hat’s OpenShift strategy. Hiring some of the most talented engineers in the industry can take years if done by a typical HR team sourcing talent. With an acquisition, all that talent becomes immediately available to Red Hat to put to use in shaping its Kubernetes solution. Speed is critical in this fast-changing Kubernetes landscape, and Red Hat wants to waste no time.
With the acquisition now successfully completed, Red Hat will be busy integrating the teams, products, and soon announcing a unified solution that is superior to what it offered before. There is ripe opportunity for it to take the best of OpenShift and Tectonic, and the various other complementary tools that both have built and gain an edge. This is easier said than done, and often culture can get in the way of these type of complex integrations. What’s clear is that the Kubernetes space is full of opportunity, and an acquisition like this promises to come out with unique innovations that moves the industry forward and capitalizes on this opportunity.
Photo credit: Pixabay