Docker is an open platform for developing, shipping, and running applications in containers.
Containers allow developers to package applications with all dependencies, enabling consistent behavior across environments.

• Fast delivery and automation – Ideal for CI/CD pipelines: from development to production, everything runs in the same container.

• Portability and flexibility – Containers run consistently across environments (dev, staging, production, cloud, on-prem).

• Efficient use of resources – Containers are lightweight and can run more workloads on the same hardware compared to virtual machines.

Docker follows a client-server model:

• Docker Daemon (dockerd) – Listens to API requests and manages objects like containers, images, volumes, and networks.

• Docker Client (docker) – The main user interface, which communicates with the daemon via CLI commands like docker run, docker build, etc.

• Docker clients can connect to remote Docker daemons over a network using a REST API.

Available for Windows, macOS, and Linux. It includes:

• Docker CLI

• Docker daemon

• Docker Compose

• Docker Content Trust

• Kubernetes (optional)

• Credential helper

A registry stores Docker images.

• Docker Hub is the default public registry.

• You can also set up a private registry.

• Images are uploaded and downloaded using docker push and docker pull.

Images – Read-only templates with instructions to create containers. Built from Dockerfile, each command creates a layer.

Containers – Runnable instances of images. Created with docker run. They include:

• Writable layer on top of the image
• Networking and process isolation
• Destroyed or persisted depending on configuration

Docker uses Linux kernel features:

• Namespaces – For isolation (processes, networking, filesystems)

• Control Groups (cgroups) – For resource limits (CPU, memory, etc.)

• On Windows/Mac, Docker runs in a lightweight VM to simulate Linux kernel behavior.

• Docker pulls the ubuntu image if it's not local.

• A new container is created from the image.

• A writable layer is added.

• The container gets a unique network ID and virtual filesystem.

• The command /bin/bash runs inside the container.

• When you exit, the container stops, but its data may persist depending on setup.