Containers are lightweight and executable application components that package up the source code with all its dependencies and operating system libraries required to run in any computing environment [3].
Since containers are small, more resource-efficient, and portable, containerized microservices or serverless functions have become very popular in cloud-native applications. Applications with a small number of containers can be easily deployed and manually managed. However, thanks to DevOps, the number of containerized applications is rapidly growing and manually managing them at scale has become impossible.
According to a Gartner report, "by 2023 more than 70% of global organizations will be running more than two containerized applications in production, up from less than 20% in 2019" [4]. The results of a survey conducted by the State of Container and Kubernetes Security show that organizations have containerized more than half of their applications and that most of these apps are already running in production [5].
Since the number of containerized applications are growing faster than ever, what could be done to address the challenges of manually managing them? The answer is container orchestration!
What is container orchestration?
Container orchestration automates the provisioning, deployment, networking, scaling, availability, and lifecycle management of containers [2]. According to a research conducted by IBM, 70% of the developers using containers also use the container orchestration solution.
Container orchestration might not be the best approach for every application. Survey respondents of the research conducted by IBM, mentioned that they do not use orchestration when scalability is not an issue, the application will not be deployed in multiple environments, the application includes few containers, or the applications' resource utilization is very small.
However, there are certain tasks that would take advantage of orchestration if you are using containers in your application. For example, the configuration and scheduling of containers, verification of the availability of containers, verification of which container applications are running in, scaling of containers to equally balance application workloads across infrastructure, allocation of resources between containers, and health monitoring of containers [1].
Nowadays, Kubernetes is the most popular container orchestration platform, although there are other orchestration tools such as Docker Swarm and Apache Mesos.
How are orchestrated solutions bringing business value?
Most respondents of the research conducted by IBM, think that orchestration provides robust security and increases productivity. Furthermore, some people think that it is easily portable and it minimizes human error, it minimizes costs by optimizing the resource usage, and it reduces the risks of downtime with deployments and rollbacks. Finally, it auto-scales to fit the needs of each application and it supports hybrid cloud environments.
How does container orchestration work?
Even though the methodologies to orchestrate can vary across tools, there are three basic steps that involve container orchestration overall.
- Most container orchestration tools support a declarative configuration model. That is, the developer writes a configuration file that defines the configuration state, and the orchestration tool will run the configuration file to achieve the desired configuration. Normally, a configuration file defines which containers will be used for the application and in which registry they are located, gives storage and other resources to containers, defines the network connections between containers, and specifies versioning.
- The orchestration tool will then schedule the containers' deployment to a host by choosing the best host based on the availability of the CPU, memory and other constraints specified in the configuration file defined in Step 1.
- Once the containers are deployed, the orchestration tool will manage the lifecycle of the containerized application. This is often done by a definition written in a Dockerfile, such as load balancing, resource allocation, ensuring the availability and performance of the containers etc.
TLDR-summary:
Although container orchestration is mostly used to reduce the effort and complexity of managing a large containerized application, its main benefit is automation! By automating operations, development teams can release new releases faster. Besides that, by orchestrating the application, companies may reduce costs due to the efficient maximization of computing resources, automated health monitoring, and relocation of containers in order to maximize availability.
About the author:
Isabella Ferreira is an Ambassador at TARS Foundation, a cloud-native open-source microservice foundation under the Linux Foundation.
References:
[1] https://avinetworks.com/glossary/container-orchestration/
[2] https://www.ibm.com/cloud/learn/container-orchestration
[3] https://www.docker.com/resources/what-container
[4] https://www.itopstimes.com/contain/enterprise-container-strategy-its-time-to-jump-on-board/
[5] https://security.stackrox.com/state-of-kubernetes-security-2021.html?Source=Website&LSource=Website
