Docker orchestration was a hot topic at DockerCon 2016 where thousands of DevOps teams gathered in Seattle last week to learn more about containerizing workloads for more agile and efficient application development.
We talked to hundreds of people during the show, and conducted a short survey in our booth to take a quick pulse of this DevOps crowd on use cases for containers, and how they use Docker orchestration frameworks. Of the more than 300 people who responded, 86% said they were responsible for building or running apps. If they weren’t, we excluded them from the rest of the survey.
We asked about use cases…and this is where I was a bit surprised. The leading use case, cited by 78% of respondents, was to use containers to deploy web services using the microservice pattern.
However, 65% of respondents said they use containers as “lightweight VMs.“ In other words, they run existing (not truly microservices-based) applications in containers instead of VMs. Big Data workloads were chosen as the third major use case (27%). Of those that cited “other,” CI/CD and testing were the most frequently mentioned use cases.
Docker’s announcements in release 1.12 at DockerCon16 provided a lot of fuel for Docker orchestrations conversation during the show. We asked respondents how they use frameworks for Docker orchestration. As expected, Docker Swarm+Compose lead in terms of current use (this was Dockercon, after all!). Interestingly, Kubernetes and Swarm tools were neck and neck in terms of interest and active research. Mesos lagged across the board. It is worth noting that this survey didn’t measure the scale and sophistication of deployments, so it’s possible that in larger scale and more sophisticated deployments, the distribution of usage will be different.
Thanks to everyone who participated in the survey. If you’re using these technologies or are in the community, I’d like to hear your thoughts about this data.