Docker - Architecture

Docker - Architecture

Understanding Docker’s architecture helps clarify how it enables containerization and differs from traditional virtualization.

Traditional Virtualization Architecture

In traditional virtualization, the architecture typically involves these layers:

1. Infrastructure: The physical hardware (Server: CPU, RAM, Disk, Network).
2. Host Operating System: The main OS installed on the hardware (e.g., Windows Server, Linux).
3. Hypervisor: Software (like VMware ESXi, Microsoft Hyper-V, KVM) that creates and manages Virtual Machines (VMs). It abstracts the hardware.
4. Guest Operating System: Each VM runs a full, independent copy of an operating system (e.g., multiple instances of Linux or Windows). Each Guest OS has its own kernel, libraries, and binaries.
5. Applications: Applications run on top of their respective Guest OS within each VM.

This approach provides strong isolation but incurs overhead due to running multiple full operating systems.

Docker Containerization Architecture

Docker utilizes OS-level virtualization (containerization) with a different architecture:

1. Infrastructure: The physical hardware (Server: CPU, RAM, Disk, Network).
2. Host Operating System: The main OS installed on the hardware (e.g., Linux, Windows).
3. Docker Engine: The core Docker software that runs on the Host OS. It includes the Docker daemon (server) and a client interface (CLI, API). The daemon manages containers, images, networks, and volumes.
4. Containers: Instead of full VMs, Docker runs applications in containers. Crucially, all containers share the Host OS kernel. Each container only packages its application code, runtime (e.g., Node.js, Python, Java), system tools, libraries, and settings needed beyond the base kernel. They run as isolated processes within the Host OS userspace.

This architecture offers significant advantages:

• Efficiency: Containers are much more lightweight than VMs as they don’t need a separate OS kernel. They consume fewer resources (CPU, RAM, disk space).
• Speed: Containers start almost instantly because they are just processes on the host OS, whereas VMs need to boot a full operating system.
• Density: You can run many more containers than VMs on the same hardware due to lower overhead.

Docker Engine Components

The Docker Engine itself follows a client-server architecture:

• Docker Daemon (dockerd): A persistent background process that listens for Docker API requests and manages Docker objects like images, containers, networks, and volumes.
• REST API: Specifies interfaces that programs can use to talk to the daemon and instruct it what to do.
• Docker CLI Client (docker): The primary way users interact with Docker. The client sends commands (like docker run, docker build) to the dockerd daemon via the REST API.

The client can communicate with the daemon on the same system or connect to a remote Docker daemon.