By Synthesis Juan Roriston and Rodney Ellis
Every company wants to accelerate business value. Kubernetes is one of many answers. But to those in the non-technical team, this may sound like just another buzzword. Below we break it down so the next time the word is mentioned, you can have an impactful say in the conversation.
What is Kubernetes?
Kubernetes was developed by Google. I announced it in 2015, with version 1.0 being released July 21, 2015. It has seen multiple versions since then and is currently on 1.21.
Kubernetes is an open-source platform that automates Linux container operations. It eliminates many of the manual processes involved in deploying and scaling containerised applications.
How and why does it drive innovation?
With most of the container management being handled by Kubernetes, businesses can focus on developing features for their application rather than worrying about its deployment. Manual configuration of container deployments carries a lot of overhead in terms of networking, security, resource allocation and application availability. By automating this using Kubernetes, you can cut down on the engineering time it takes to deploy workloads. This will help developers focus on accelerating business value.
What are its benefits?
Kubernetes can seamlessly integrate with CI/CD tools allowing for quicker deployments and consistent releases. It is also very flexible and portable meaning that you will not be locked into a cloud provider or containerisation tool. Kubernetes is free to use and is still supported by the open-source community to this day. It also has self-healing functionality built into it and will try and correct any faults and failures that the underlying infrastructure may experience. Using some of the native built-in Kubernetes functionality for example HPA (Horizontal Pod Autoscaling) will automatically scale workloads to meet the demands of the application.
Is this relevant for every application?
No, Kubernetes is a great tool, but it is just that, a tool like any other. Finding the right application to host on Kubernetes can be tricky as it was not made to accommodate every application’s architecture type. Having a monolithic application architecture or a small-scale application that can be easily hosted on serverless should disqualify it from being hosted on the platform.
Is this relevant for every size company?
Yes, the size and industry of the company are irrelevant when deciding on whether or not to use Kubernetes but it is rather the application itself that will determine whether or not you will need to use it.
How do you know if your organisation should be using it?
According to a 2019 Cloud Native Computing Foundation survey, 84% of respondents said that they were running containers in production. If your business falls into this bracket, then you should be considering Kubernetes to manage your workload.
The other considerations to make are the application itself and if it is viable to be hosted on a Kubernetes cluster. Kubernetes can make use of other tools such as Kafka for streaming data.
What are the alternatives?
You can manage your container deployments manually, but this takes a lot of time and effort to achieve and manual intervention often leads to human error.
Other alternatives will be using other orchestration tools such as Docker Swarm or AWS Elastic Container Service although these are mostly used for simple and small applications. Kubernetes is needed for more complex workloads.
Is it expensive?
Kubernetes is open source and free to use. The cost you will incur will be on the underlying infrastructure needed to host the cluster. This can range in price as the cluster will need to meet the resource demands of the application.
Examples of where it improve businesses or create tangible benefits?
At customers, we have seen an increase in deployments and feature releases to staging and QA environments which allow for features to be deployed to production much faster than before. This greatly improves the business value. We have also seen fewer interruptions to application and service availability. We have also lowered cost by implementing scaling features such as HPA to decrease the infrastructure needed during times of low usage.
