“Companies are realizing the productivity and flexibility gains they were expecting, and use of container technology is clearly on the rise,” said Shippable CEO Avi Cavale in a statement.
Those are optimistic words, and almost 90 percent seems like a strong vote of confidence for containers such as Docker and others, validating the intense hype the technology has generated in the last couple years. Perhaps that’s why more than half of developers (52 percent) told Shippable, which sells a continuous delivery platform, they already use containers in production, at least for new applications.
Modest gains in development speed
Other numbers from Shippable’s survey paint a more complicated picture, however. For example, for almost three quarters (74 percent) of respondents, containers have allowed them to ship new software faster. That’s great, right? But those gains appear surprisingly modest, merely 10 percent. Only 8 percent of respondents report that containers enable them to ship apps 50 percent faster.
Similarly, only a third of developers said container technology makes release cycles “much faster,” and a measly 17 percent—about one in six—said containers help them do the same amount of work with fewer coders. Those kinds of numbers are simply not enough to justify the investment in time and effort required to implement containers.
Lingering container objections
Why the lackluster results? Some of it is due to perceptions that container technology remains difficult, immature and unproven. Other developers have had issues making their existing infrastructure work with containers, or they don’t think they have the required skills. Security can also be an issue.
But it’s likely a chicken-and-egg issue. ROI on container technology can’t be established until it’s widely deployed, and it’s hard to earn wide deployment until you can show a clear ROI.
Not surprisingly, when container deployments do come, they mostly happen in the cloud. The biggest group of container users (31 percent) run their containers in the public cloud infrastructure, followed closely by private implementations (30 percent) and hybrid environments (17 percent). On-premise users apparently weren’t invited to the container party—only 2 percent of developers said they run containers in on-premise data centers.
Despite Amazon Web Services’ (AWS) overall cloud dominance, however, only 43 percent of survey respondents run their containerized apps on AWS. Google Compute Engine (52 percent) and Microsoft Azure (49 percent) were more popular choices. That result was especially unexpected, at least by me.
Overall, the survey reinforces the conventional wisdom that containers are hot, but not yet fully understood by all developers. Still, this survey looks a lot different than it would have a year ago, and next year’s container survey results are likely to be even more positive.