Learning Apache OpenWhisk

Chapter 12. Deploying OpenWhisk with Kubernetes

Apache OpenWhisk is available out of the box in the IBM Cloud and in the Adobe Cloud as I/O Runtime, but it is a true open source project and can actually be deployed everywhere. It can be installed in other clouds, either public or private, or on your own servers. In this chapter, you will learn how to install OpenWhisk both for development and production, and how to deploy complex applications in it. But before going into all the details, let’s cover some of the basics.

OpenWhisk is actually built on Docker. Docker is known as a “lightweight” virtualization environment. It provides a format for defining disk images, named Docker images, and a way to launch them.

There is a public distribution point known as Docker Hub. OpenWhisk is distributed on Docker Hub in the form of a collection of Docker images, deployed to create a “control plan.” When you deploy an action, you communicate with the OpenWhisk control plan, and it creates other containers to run actions. Runtimes to execute actions are also Docker images.

Generally, Docker alone is not enough to run complex, multiserver applications. You need another piece of software, called Kubernetes. In this chapter, we focus on running OpenWhisk in Kubernetes, either locally for development purposes, in the cloud, or on a bare metal server.

Note

Docker is essentially a Linux product, but a version for Windows and Mac (called Docker Desktop) exists. Docker Desktop also includes Kubernetes. ...