Mirantis Launches Open Source Kubernetes Control Plane Project
Mirantis today launched an open source k0smotron control plane project for Kubernetes that makes it simpler to centrally manage multiple clusters.
Jussi Nummelin, senior principal engineer for Mirantis, said the core software would also be used by Mirantis to bring the control plane it created for Kubernetes in line with the cluster application programming interfaces (APIs) defined by the technical oversight committee for Kubernetes.
The goal is to let k0smotron automatically provision underlying virtual machines (VMs) and networks using the cluster API, he noted.
The Kubernetes TOC has thus far shied away from trying to standardize control planes, instead allowing vendors to add value atop the APIs it defines. Mirantis is now laying the foundation for a more centralized approach to managing Kubernetes clusters at scale via an open source project that includes capabilities such as support for remote worker nodes.
It’s not clear how much support there is for k0smotron just yet, but each vendor’s efforts to build their own control plane are redundant. In theory, k0smotron creates an opportunity to streamline operational overhead in environments where multiple Kubernetes control planes might reside.
That’s becoming an issue as more Kubernetes clusters are deployed everywhere from the network edge to the cloud. IT teams are looking for ways to streamline management overhead.
Of course, many of those same IT teams are waiting to see how artificial intelligence (AI) might be applied to achieve that goal. There are multiple open source projects that are applying AI to Kubernetes management, but it’s still early days in terms of how AI for IT operations (AIOps) for Kubernetes might be applied in a production environment.
In the meantime, as more tools become available that abstract away the inherent complexity of Kubernetes clusters, it’s clear that IT teams are becoming less intimated by a platform that is simultaneously one of the most powerful and challenging to manage ever to be adopted in mainstream IT environments. The issue, as always, is there is not enough Kubernetes expertise available, so increased adoption of the platform depends heavily on the graphical tools made accessible to IT administrators versus always having to be dependent on a DevOps engineer that is hard to find and retain.
Eventually, the increased pace at which cloud-native applications that run on Kubernetes clusters are developed will force the Kubernetes management issue. In theory, those applications could be centrally managed on a handful of large clusters, but in practice, IT teams are deploying smaller Kubernetes clusters that are dedicated to a handful of applications. That approach provides greater isolation for applications that are often funded by different departments within the same organization.
Regardless of the approach to managing Kubernetes, it’s clear that many organizations may soon find themselves managing hundreds of Kubernetes clusters. The challenge is that as the number of clusters continues to expand, there may not be a corresponding increase in the number of IT professionals hired to manage them.