Originally posted on CA Technologies Highlight
Observations from DockerCon, KubeCon and Container World
Docker is to containers what Kleenex is to tissue. When someone talks about containers, it’s a safe bet they are talking about Docker containers. This is good news as we can all agree on the basic building blocks for our Microservice applications. However, we are running applications – not building blocks. Container orchestration is where the majority of the recent power shift has occurred.
The container orchestration players
Apache Mesos, Kubernetes, OpenShift, Docker Swarm, and Rancher were some of the early players in the container orchestration space. Each had a niche and served it well. However, over the last year, everything has shifted towards Kubernetes. OpenShift 2 was a proprietary orchestrator that adopted Kubernetes in version 3. Recently, Rancher announced their 2.0 product would use Kubernetes as a base instead of their original orchestrator. And most recently of all, Docker added support in their tools for Kubernetes.
Here are a few reasons for Kubernetes’ rise to prominence:
- The Cloud Native Computing Foundation (CNCF) governs the development of Kubernetes. This means there isn’t a single commercial entity controlling its direction. Apache Mesos is similar, however, it may have been ahead of its time.
- Kubernetes came from the Borg, which came from Google. The allure of running your container application the same way Google runs theirs is enticing, whether or not you need all the same capabilities (and complexities).
- As companies and vendors consolidate on a tool, it tends to bring others along. As more and more companies merged on Kubernetes, the more dominant it became.
- Kubernetes is great software.
You’ll notice that simplicity is not one of the listed traits.
At KubeCon’17, there was plenty of talk about making the container orchestration layer “boring”. However, this will take time and significant tooling. For now, you will find yourself writing more YAML than ever.
Takeaways from industry conferences
Over the past year, I had the opportunity to attend DockerCon (April), KubeCon (December) and Container World (February). It was extremely interesting to witness the conversation around orchestration evolve over the course of just 10 months! Kubernetes’ growing buzz from the Spring to the Winter was impossible to ignore.
At DockerCon ’17, much of the focus was about containerizing new and existing applications, and getting them into production using Docker Swarm.
The dialogue at KubeCon ’17 centered around container applications needing ‘ilities such as secure-ability, scalability, observability, and maintainability. We learned how companies and opensource projects are helping enhance the Kubernetes ecosystem (Istio, for instance, is a great example of an opensource project enhancing Kubernetes).
At Container World ’18, the discussion was a combination of adopting containers, running and supporting containerized applications and how to get started. When it came to orchestration, Kubernetes dominated the conversations.
It was expected for a Docker conference to primarily talk Docker solutions, and a Kubernetes conference to talk Kubernetes solutions. It would be logical for a neutral conference to discuss both. However, that clearly wasn’t the case at Container World.
So, what’s next?
Will Kubernetes be synonymous with container orchestration just as Docker is synonymous with containers? Most signs are pointing to yes. And I believe DockerCon ’18 might just provide the answer.