Docker is mainstream. Kubernetes is everywhere. Where does the container ecosystem go from here? Keep reading for a look at what’s next in the world of containers and microservices.
Just a few years ago, the question of whether you should use containers at all was still a matter of debate.
Then, the conversation shifted to how much of your workloads should be migrated to containers, and which workloads are better served by virtual machines.
Today, whether you should use Docker is no longer in question. Nor is there much room left for debate about which container orchestrator is best; in most people’s minds, Kubernetes has won that war.
Still, the world of containers is not done evolving. The following topics will drive innovation in the container ecosystem in 2018 and beyond.
Serverless vs. Docker
One issue that is likely to help shape the container ecosystem going forward is the question of whether serverless computing is a better fit for most workloads than Docker.
I tend to think that is a silly debate, because the functional overlap between Docker and serverless is limited.
Still, there are folks out there who see serverless computing as a better solution to their IT challenges than Docker.
System container technology has been around since long before anyone heard of Docker.
However, the vast majority of growth in the container ecosystem over the past few years has centered on application containers, which are the primary use case for Docker.
There are signs, however, that interest in system containers by major vendors including Red Hat is growing.
Expect system containers to represent another area of growth for the container world in the new year.
Service mesh is a term that is suddenly coming into vogue in the world of containers and microservices.
There is no precise definition of the term (although the one from the Buoyant/Linkerd folks is probably the most common). It means somewhat different things to different people.
In general, however, a service mesh is the infrastructure layer that allows different microservices to communicate and interact.
Most microservices environments already have a service mesh of some sort, because you almost always need a way for your microservices to communicate.
Going forward, however, service meshes are likely to become a more deliberate and sophisticated part of a microservices architecture. They will also be easier to implement, thanks to entire software platforms created for the purpose, such as Linkerd.