Docker builds are fast because of the build cache: when you rebuild an image, Docker reuses the results of instructions whose inputs have not changed, rebuilding only what is necessary. Understanding how this cache works — and how it is invalidated — lets you write Dockerfiles that rebuild in seconds instead of minutes, a major productivity and CI-speed win.
The cache is layer-based: each instruction produces a cached layer keyed by the instruction and its inputs. If an instruction and its inputs are unchanged, Docker reuses the cached layer; if something changes, that layer and every layer after it are rebuilt. This cascading invalidation is the single most important thing to understand, because it dictates the optimal order of instructions.
This lesson covers how the build cache works, why instruction order matters so much, the canonical pattern of copying dependency manifests before source, keeping the build context small with .dockerignore, and reducing layers. These optimisations turn slow, wasteful builds into fast, efficient ones.