Docker is now about 6.5 years old. But evidence suggests that only a minority of production workloads are containerized. Here’s a look at why container adoption is relatively low and what needs to change for more companies to adopt containers.
Adoption Rates in 2019
First, though, let’s examine what container adoption rates look like currently.
One of the most recent studies of containers’ market share is a Red Hat survey from earlier this year. Red Hat found that 57 percent of the companies surveyed were using containers. That figure aligns broadly with the findings of another recent study by DivvyCloud, which reported that 78 percent of companies are either using containers already or plan to implement them in 2019.
But when you look more closely, the numbers surrounding adoption rates are less impressive. A mere 13 percent had containerized at least half of their workloads, according to Red Hat.
Meanwhile, the fact that DivvyCloud chose to lump together the number of companies that had or were planning to adopt containers (as opposed to reporting on how many had already adopted them) suggests that the rate of actual container adoption is on the lower side. (It’s in DivvyCloud’s interest to play up the rate of adoption since the company provides security services for containerized infrastructure.)
In short, it appears that while the vast majority of companies are interested in containers, the rate at which they are migrating to containerized infrastructure remains low.
Challenges to Container Adoption
Here are four reasons that explain containers’ relatively low impact on the real world:
- Lack of staff expertise. Red Hat mentioned at the time of the release of its container adoption survey that finding IT staff who have the skills to work with containers is one barrier to adoption rates. That’s a bit surprising given that Docker is no longer a radically new technology. But I suppose it’s still new enough that it’s not something most IT engineers or developers have worked with previously unless they had a specific reason to do so.
- Lack of confidence in container performance. According to Red Hat’s survey, almost half of companies are unsure whether containers perform better than virtual machines. Another 12 percent believe containers perform worse than virtual machines. This data suggests that many organizations lack confidence in the oft-touted performance benefits of containers. That’s not unfair; the performance advantages of containers are less impressive than many people assume, even though containers are generally more efficient than virtual machines.
- Difficulty choosing how to deploy containers. The container ecosystem is broad and diverse. There are several major orchestrators and about a dozen container runtimes. Choice is great in many respects, but it can be a drawback for deciding how to deploy containers. I’m betting that some companies are having a hard time jumping on the container bandwagon because they can’t decide which tools to adopt to fill in all of the layers of containerized infrastructure.
- Security. No matter how you slice it, containers are more complex than virtual machines. By extension, it’s easier to end up with security gaps in containerized infrastructure. Plus, the rash of headlines over the past several years about security flaws in Docker itself have no doubt given some companies pause about the security implications of container adoption.
It’s clear that lots of organizations are interested in containers and want to adopt containers. But a variety of hurdles continue to prevent many of them from making the switch to containers or from containerizing as many of their applications as they would like. These hurdles can be overcome, but they present real barriers to container adoption, which container advocates often overlook amid all the hype about the many benefits of containers.