DockerCon 2016 has come and gone. What does Docker need to do between now and the next DockerCon to assure the success of Docker containers in the enterprise? Here’s a look at where Docker stands currently and the challenges that remain for it to solve.
To be sure, Docker has come far since its beginnings as just another obscure Linux containers project at dotCloud in France. But it still has not fully convinced enterprises that containers are ready for production environments. Nor has it stemmed competition from other container providers, including CoreOS.
To keep moving forward, Docker needs to do the following:
Continue Enterprise Adoption of Docker Containers
However, the way these surveys define “production” is hazy. If you read them closely, it’s easy to come away with the sense that most of the people currently using Docker in “production” are actually just internal development teams. There have been few stories so far of major companies containerizing large portions of their infrastructure.
To maintain momentum, Docker needs to keep pushing enterprises not just to use containers for their development workflows, but also to replace their virtual servers with them. Otherwise, Docker will remain just a tool that programmers like, while clouds and data centers continue to run on VMware.
Deliver Solid Persistent Storage
Docker Data Volumes are Docker’s current solution for persistent data storage on containers. They do the job, but not in an especially simple or flexible way.Third-party vendors are delivering better options. Most recently, Red Hat announced plans for a scale-out container storage solution based on Gluster and OpenShift. That’s good for Docker because it adds another storage solution to the ecosystem. But storage options such as Red Hat’s are tied to particular vendors’ own platforms. They’re not all-around answers.
To solve this challenge, Docker needs to offer better platform-agnostic storage solutions that anyone can use in any environment. It should not keep relying solely on partners to fill in the cracks in the container storage stack.
Define Its Channel Vision
Speaking of partners, Docker also needs to define its partnership positions better. The company announced at DockerCon that it is going to bake orchestration tools into Docker itself. That makes Docker a much more complete solution for container deployment and management. It also undercuts the value of offerings from Docker partners such as Rancher, Aqua and Codefresh, which deliver orchestration tools for containers. That means those partners are likely to become isolated. If Docker wants to avoid burning the partners that have helped build the container ecosystem up to its current state, it needs to offer them new ways to keep integrating, reselling and adding value to Docker containers.
There are still areas ripe for development; for instance, my colleague Chris Riley points to rich visualizations as one area where partners can still build significantly upon Docker.
If Docker doesn’t encourage partner engagement such as this, the ecosystem is likely to become stale. And Docker is not Apple: It can’t build an entire ecosystem and software stack without help from the channel.
Beat Back Competition
For all the headlines Docker has made, its competitors are still thriving. That’s a good thing for the container ecosystem as a whole. But for Docker, having the likes of CoreOS (which builds an alternative container platform) and Canonical (which builds LXD and claims it is not trying to compete with Docker, but easily could) perennially waiting in the wings is not a good thing.
Docker doesn’t need to obliterate the competition. It probably can’t, in any case. But the pressure from competitors is still stiff enough that Docker’s own future is not entirely assured, and the company needs to keep working to address that.