DockerCon 2017 has come and gone. What did it tell us about the current state of Docker, Docker containers and the ecosystem? Keep reading for an overview.
First of all, this year’s DockerCon provided an indication of just how massively popular Docker containers have become in a few short years. Attendance, which was around 5,500, increased by a factor of 10 from the 2014 DockerCon.
A Strong but Boring Ecosystem?
Yet while there’s no question that the Docker ecosystem is robust, I was struck by how boring this year’s DockerCon was.
OK. I’m being a bit provocative here. It’s not as if nothing exciting happened at DockerCon. There were some relevant announcements, such as:
- The launch of LinuxKit, a platform for building customized Linux-based operating systems for hosting containers.
- Docker’s launch of an “application modernization program,” which helps clients redesign their software architectures for the Docker age.
- StorageOS announced a new persistent storage solution for containers—a corner of the container ecosystem that remains quite under-saturated, especially after ClusterHQ shut down late last year.
- Tools including Yipee.io and Qubeship.io debuted to make it easier to work with microservices applications.
But apart perhaps from the LinuxKit news, none of the above amounts to truly momentous innovation. The Docker ecosystem is filling in around the edges a bit, but its core is no longer changing.
In that sense, this year’s DockerCon was a major departure from last year’s. At DockerCon 2016, Docker announced that it was baking Swarm into the core Docker platform.
The move had major repercussions for the ecosystem for two reasons: Including Swarm with Docker ensures that developers working with Docker are ready to use Docker at scale out-of-the-box, and—most important—Swarm’s integration upped the ante for third-party vendors working on container orchestration or management tools of their own.
This year’s DockerCon saw no huge changes to the Docker platform itself. And while LinuxKit is cool, it’s not revolutionary. Building a customized, Linux-based OS for hosting a container is not a new idea. Nor is LinuxKit the only way to do it.
Probably the most interesting dimension of LinuxKit is that it could make it possible to host a containerized Linux application on a Windows host without virtualization. That would be exciting.
But that eventuality is the only Docker-related news this month that really got my heart pumping. The rest of the DockerCon announcements represented substantive but standard fare.
That leaves me wondering whether Docker has now gone mainstream to the point that the exciting days are over. Will vendor partnerships, product integrations and the occasional open-sourcing of proprietary tools be all we have to look forward to in the future of the Docker ecosystem? I think it may be.