How Docker Image-Container Dependency and Client-Server Architecture Work

Docker images are built from read-only layers, each corresponding to a line in a Dockerfile, and cannot be deleted while a container is actively using them. When a container is created, it does not copy the image but instead holds references to its layers, along with its own writable layer for storing runtime changes. This design pattern is known as Copy-on-Write (CoW), meaning the image remains unaffected by any modifications made inside the container. Docker operates on a client-server model where the Docker client sends requests to the Docker daemon, which handles execution and fetches images from public repositories if not found locally. Tools like 'docker ps --size' and the Dive application help users monitor container sizes and inspect individual image layers respectively.