Gabi Moisa - Fotolia

What containerization technology do I need for my data center?

Containers in the data center can help maximize server capacity, but there are a few things to consider before a Docker or Kubernetes implementation.

Containerization technology includes all the necessary components to run a data center application, but it requires less storage and is easier to install on servers than traditional software packages.

To determine how to deploy containers, you must first identify the container type. It's likely that the container that comes with software is a Docker container because Docker is a major player in the container market. You may also encounter rkt and systemd depending on the container application and system requirements.

Rkt is a part of CoreOS and is completely open source. Systemd containers don't offer all of Docker's features, but they are easy to implement and are automatically installed in the Linux OS. Systemd provides socket activation and nspawn containers, which enable you to run more than one process within a container.

Support container networks

You don't need much if you want to run a few Docker containers, and it's best to use the Linux server where you installed the container engine. This daemon process runs on top of the Linux kernel and allows you to run containers on a Linux box.

Docker Enterprise edition is available for all Linux types and is installed with distribution repositories. Running just the container engine doesn't provide any redundancy, however, and is recommended only for small and noncritical workloads.

If you want some scalability and redundancy in your containerization technology, you'll need container orchestration tools. An orchestration setup may consist of CoreOS, a minimized Linux operating system designed with one purpose: to run containers as efficiently as possible.

A CoreOS setup is comprised of multiple hosts that share information about any active containers and bring high availability to the server. If anything goes wrong within the server or container, another node will always be available to take over the workload.

Another well-known option for container orchestration is Kubernetes. Kubernetes runs on top of the container engine, and its purpose is to ensure you'll always have enough containers to guarantee continuous operation. You can install Kubernetes in the data center on a physical or virtual server, but you must be sure to use a few servers to ensure processing continuity and scalability.

Another alternative is to run containers outside of the data center on a private or public cloud. All major public cloud providers offer a container engine, and Kubernetes has rapidly become the standard within the public cloud. If the container needs to be available for users around the world, a public cloud is probably the best choice.

Dig Deeper on Data center ops, monitoring and management

Cloud Computing
and ESG