CSE Roadmap

Common tier · about 2 hours · 6 parts

Part 4 of 6 · 1 min read

Images and Docker

Build once, store by version, run anywhere.

BM describes Docker as a platform built on lightweight containers that does much of the work of handling them. In Docker you create and deploy "apps", the equivalent of images in the VM world. Docker handles container provisioning and some networking, and offers a registry, a store for versioned images.

The idea for a pipeline (module 4): the build stage produces an image from the code, stores it in the registry, and every later stage runs that same image. Because the image hides the technology inside, the deploy step is the same for every service. BM stresses the value of a uniform deployment interface: whatever the platform, a service should be deployable the same way every time, from the laptop to production.

Another use from BM: for development and test, Docker lets one VM run many service containers, which relieves a laptop that cannot run a whole system as separate VMs.

BM's own limit: Docker on its own works like a simple platform for one machine. Spreading services across many machines needs more.

Image and registry

An image is a frozen, versioned package of a service and everything it needs to run. A registry is a store that keeps images by name and version.

Example: checkout:2026-11-13-a1 is built once, stored in the registry, and the same bytes run in pre-production and production.

Check yourself: 1. What is the difference between an image and a container? 2. Why run the same image in every environment? 3. What does BM say Docker alone does not solve?