Docker images are organized in layers using Union File System (UnionFS). When you pull an image or build a new one, each layer is downloaded or built separately and stacked on top of each other to form the final image.
Every Docker image starts with a base layer, which is an OS image like Ubuntu, Alpine or even Windows. Each subsequent layer corresponds to a series of changes made to the file system of the preceding layer. This could be adding a new file, modifying an existing file or deleting a file.
Each of these layers is read-only, except for the top most layer, also known as the container layer, which is writable. When you create a Docker container from an image, Docker adds this read-write layer on top. Any changes made to the running container such as writing new files, modifying existing files, and deleting files, are saved in this writable layer.
Layering provides some significant benefits:
1. Layer Reusability: If multiple images share the same base, Docker will only store that base once, saving disk space and network bandwidth when pulling images.
1. Layer Versioning: Docker uses a content-addressable file system. Each layer has a unique identifier and can be shared among images, providing a version control system at the layer level.
1. Speed: Pulling layered images from a registry is much faster than pulling a single monolithic image.
1. Efficiency: It’s possible to cache layers when building an image, which can significantly speed up the build process.
Dockerfile instructions like RUN, COPY, or ADD creates a new layer in the image for each of these operations.
