Containers: Software Containerization
Introduction
Containerization is a form of operating-system-level virtualization or application-level virtualization over any kind of machine that can run the container runtime.
Basics
Containers are a standard unit of software that packages up code and all its dependencies, including the operating system, libraries, system, etc. into a computing environment that can quickly and reliably run on any machine. Container images tend to be lightweight and portable, and can be used to deploy applications on any machine under any configurable environment that native deployments can configure.
Container images become containers at runtime and in the case of container runtimes - images become containers when they are executed. Regardless of whether the runtime runs on Linux, Unix, Windows, or macOS, the container is always executed in the same way no matter the infrastructure. Containers are isolated from one another and bundle their own resources, so the applications run uniformly and their isolation provides some greater degree of separation or sanboxing from the host OS actually running the container. This provides these benefits.
- Standardization
- The container runtime ensures each container is run in the same way regardless of the host machine.
- Lightweight
- Containers share the host OS kernel.
- Thus doesn't need to run a full-blown OS on each machine.
- Thus more efficient than virtual machines.
- Secure
- Applications are safer in containers
- Can be configured to have no access to host OS resources at all.
- Can be configured to have no access to other containers.
Docker
Docker is a software stack that provides containerization of applications and virtualized operating systems. It was one of the first containerization platforms, that gained prominence in the early 2010s.
Docker in the 2020s is a mature platform that is used by many organizations. Although the landscape of containerization platforms has changed significantly, Docker is still a popular choice for many organizations despite the rise of other containerization platforms that address some of the limitations of Docker.
Networking
Containers can be connected to each other and to the outside world. Most container runtimes have features to network containers together or through the host machine in host-mode by opening ports and mapping them to any container port.
In docker, this is done with the --network flag.
For more docker-specific networking information,
see Docker: Networking.
Use Cases
Service Orchestration
There are ways to orchestrate containers either in a statically defined or dynamically defined set of configurations.
Compose File
Note: To learn more about Container Compose File Formats, read this link.
Container Composition is a way of statically defining a way that
one or more containers should be configured and ran.
This composes a lot of seperate docker commands into one relatively simple and
easy to maintain compose file.
Then to run the service, simply type docker-compose up in the terminal.
Database Containerization
There are some good use cases for containerizing a database. With some caveats that are worth noting, namely storage performance and greater difficulty in managing the database with respect to backups, migrations, and logging. For more info, read Database Containerization.