Production Kubernetes

Chapter 3. Container Runtime

Kubernetes is a container orchestrator. Yet, Kubernetes itself does not know how to create, start, and stop containers. Instead, it delegates these operations to a pluggable component called the container runtime. The container runtime is a piece of software that creates and manages containers on a cluster node. In Linux, the container runtime uses a set of kernel primitives such as control groups (cgroups) and namespaces to spawn a process from a container image. In essence, Kubernetes, and more specifically, the kubelet, works together with the container runtime to run containers.

As we discussed in Chapter 1, organizations building platforms on top of Kubernetes are faced with multiple choices. Which container runtime to use is one such choice. Choice is great as it lets you customize the platform to your needs, enabling innovation and advanced use cases that might otherwise not be possible. However, given the fundamental nature of a container runtime, why does Kubernetes not provide an implementation? Why does it choose to provide a pluggable interface and offload the responsibility to another component?

To answer these questions, we will look back and briefly review the history of containers and how we got here. We will first discuss the advent of containers and how they changed the software development landscape. After all, Kubernetes would probably not exist without them. We will then discuss the Open Container Initiative (OCI), which arose ...