What Are Containers?
What a Container Is and How It Packages an Application
What is a container? It is an isolated environment that runs an application together with the libraries and settings it needs, using the operating system kernel of the host it runs on. In plain terms, a container is a way to package software so that it runs the same way wherever it is started.
The package is called a container image. It holds the application files, the runtime it depends on (for example, a Python or Node.js interpreter), libraries and configuration. Unlike a virtual machine image, it does not include a full operating system kernel. The kernel is the core part of an operating system that manages hardware and processes, and containers borrow it from the host. Tools such as Docker build and run these images; see what Docker is.
When a container starts, the host's kernel creates an isolated set of processes for the application. The app sees its own file system, network and list of processes, but those processes are ordinary processes running on the host.
The kernel features behind this separation are explained in cgroups and Linux namespaces.
Why Containers Start Quickly and Run Consistently
Containers start fast because there is no operating system to boot. The host kernel is already running, so starting a container is closer to launching a program than to powering on a computer. Containers can be created, stopped and replaced in seconds, which helps when an application must scale up under load.
Consistency is the other benefit. A developer can build an image on a laptop and run the same image on a test server or in production. Because the dependencies travel with the application, the common problem of software that works on one machine but not another becomes less frequent.
Containers have limits. They share the host kernel, so they only run software compatible with that kernel. A Linux container needs a Linux kernel, which is why Windows and macOS users run Linux containers inside a lightweight Linux virtual machine. For a side-by-side view, see containers vs virtual machines.
Example: One Website Split Into Several Containers
A hosting company runs a content management system as several containers: one for the web server, one for the PHP runtime and one for a cache. If the cache needs an update, the company replaces only that container's image and leaves the others running. Each container can also be given limits on the CPU and memory it may use, so one busy component cannot starve the rest.
Frequently asked questions
Are containers the same as virtual machines?
No. A virtual machine runs a complete guest operating system with its own kernel on top of a hypervisor. A container shares the host's kernel and packages only the application and its dependencies. Containers are lighter and start faster, while virtual machines offer a stronger boundary between workloads.
Do containers keep their data when they are removed?
No. Data written inside a container is lost when that container is removed. Data that must persist is kept in a volume or in a directory on the host, which exists independently of the container and can be attached to a new container.