Containers

A container is a lightweight, standalone unit of software that packages:

• application code,
• the runtime platform (e.g. Python, Node.js, Java),
• system tools and libraries used by the application,
• configuration defaults.

All of this is bundled in a container image, so the application can run the same way in different environments (developer laptop, test server, production cluster, cloud provider X or Y).

Key idea:

“If it runs in my container, it will run the same anywhere this container engine is available.”

Common container engines: Docker, containerd, CRI-O, etc.

The difference with Virtual Machines is the virtualisation layers:

• Each VM has a full guest OS
• A VM hypervisor can emulate different hardware platforms
• Start slower and use more resources

Container engines

• Share the same host OS kernel
• Depend on the hardware the host OS runs on (hardware/CPU architecture)
• Start fast and use less resources

Starting Points

• Docker tutorial
• Kubernetes
• Container migration

Key Points

• You can build docker containers and test them locally
• you can deploy docker containers to the cloud
• you can deploy a multi-docker container to a (cloud) kubernetes cluster