liveBook · Manning

Container

As explained in Machine Learning Platform Engineering

A software container is an isolated, portable execution unit that packages an application with its dependencies.

A container packages an application with its dependencies so it can run consistently across computing environments. A Docker image is the package; a container is a runnable instance of that image. section 3.2 section 3.2.3

A practical Docker workflow moves an application from local development toward deployment:

  • Write the application and list its dependencies, then verify it runs locally. section 3.2.1
  • Describe packaging instructions in a Dockerfile and build a tagged image. section 3.2.3
  • Run the image, map a container port to a host port, and check that the service responds. section 3.2.3
  • Push the tagged image to a container registry so a deployment environment can retrieve it. section 3.2.3
Figure 3.3 Process of building the Docker image from the local desktop and pushing it to the container registry, which is where a server retrieves the image and runs it as a container
A diagram of a container register

AI-generated content may be incorrect.

Follow the image from a local build to a registry, where a server retrieves it and runs it as a container. section 3.2.3

How Machine Learning Platform Engineering covers it

How deep it goes
Defines containerization, then walks through a FastAPI app, image building, port mapping, tagging, and registry push.
Examples
Uses a FastAPI random-joke service and a containerized YOLOv8 Bento.
What you can do after
Build and run a Docker image, map its service port, and push a tagged image to a registry.
How it connects
Connects Docker images and registries with Kubernetes, CI/CD, and ML service deployment.

Also explained in

Compare how 27 books explain container →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning Platform Engineering. The section links point to the parts of the book it draws on.