Docker is dead

To say that 2017 was a tough year for Docker would be an understatement. I can not remember any startup from the Valley that was so well-known and well-supplied with money, except Uber, who would have spent 2017 as indistinctly as Docker. 2017 was the year Docker, a great piece of software, was completely destroyed by bad business management. This article - a side view of how and where Docker went wrong, and how attempts to fix something turned out to be too weak and belated.

Docker - good software

Honestly, Docker helped revolutionize software development. To base on such basic Linux things as control groups, namespaces, process isolation, etc. and combining them in one tool is a terrific achievement. In 2012, I tried to figure out how to make the development environment more portable. The growing popularity of Docker allowed to turn the development environment into a simple, versionable Docker file. The tools used from Packer, Vagrant, VirtualBox and tons of strapping were shrunk to just one Docker. Docker UI is really good too! This is a good tool with many applications. The guys from the Docker team should be proud of the tools they have been able to build.

Docker is the favorite of the Valley.

Docker's early success helped the company build a large community around its product. The same early success allowed to raise funding round after round. Well-known investors, such as Goldman Sachs, Greylock Partners, Sequoia Capital and Insight Venture lined up with Docker money dumpers . To date, Docker has raised investments totaling from $ 242 to $ 250 million.

But, like many well-funded startups of the 2010s, Docker made a number of personnel errors. In the process of growing Docker, some lousy people remained unraveled at the company. And this was the reason for my personal dislike for the leadership of this company. The product is still of high quality, but it is absolutely not excuse the behavior of the company. Unfortunately, this is a fairly common situation among the pets of the Valley, and this should no longer be the case.

Kubernetes hurts Docker.

Docker's unenviable fate was brought about by the growth of Kubernetes. Docker did not make any progress towards supporting Kubernetes, favored by the free container software community. All the company's attention was given to a competing product - Docker Swarm. This decision was made in spite of the fact that Kubernetes placed the focus on containers based on Docker. By the way, Docker Captains at the beginning of the year confirmed that Kubernetes’s discussions in articles, meetings and conferences were not approved by the company.

And on dockercon17 this anti-Kubernetes mantra was repeated. And then, suddenly, on dockercon EU 17 Docker decided to put everything on Kubernetes . A sudden change of mood was apparently associated with a steady rise in Kubernetes dominance. And this situation was supported by the fact that Docker became a sponsor and participant at KubeCon + CloudNativeCon North America 2017

Moby?

No one understood what Docker meant when he announced Moby on dockercon17 in April. Moby was introduced as a new main branch of the Docker project. But the release of Moby was not announced in advance. Millions of voices cried out in horror when on the GitHub the Docker suddenly changed to Moby ( https://github.com/moby/moby/pull/32691 ) when Solomon Hikes spoke at dockercon17. So abrupt and poorly thought out changes even required direct intervention by GitHub employees .

And not only the changes themselves were badly thought out, but also the announcements of these changes. Then I had to apologize and explain on my fingers what exactly had happened. All of this even more thickens the clouds on the containerization and Docker ecosystems on the cloudy sky of the so-overcast sky (or now Moby?). Some still haven't moved away from the way Docker was rolling out Moby. The Docker brand suffered from this, of course, too.

Cold hugs for Kubernetes

The belated and awkward acceptance of Docker Kubernetes at the last possible moment is a sign of impending fall. When Solomon Haykes was asked if Docker Swarm was alive, he replied (on Twitter): "The Docker will continue to support both Kubernetes and Swarm at the highest level and will stimulate their mutual development. Openness and choice create a healthier ecosystem for everyone. " The problem here is Docker Swarm is not well supported. Not at all. The Docker Swarm development team and a handful of their open source developers will not be able to compete with the Kubernetes community. No matter how good the Docker UI is, Kubernetes UI is much better. It’s almost as if Docker had dropped to the status of a marginal consulting firm in the world of containers.

Conclusion

Docker's main problem is the lack of a clear leadership course. It seems that the strategy rested on one particular person. And he was pushed further and further away from the center of the company, although he is still in it. The company restructured and focused on supporting the enterprise. This shift is justified from the point of view of investors (the company has fiduciary obligations, after all). But this shift has definitely reduced the coolness of the brand, which initially fed on its wild success. As the saying goes, "no one can kill a great civilization. Except for it itself." And that's exactly what happened with Docker.

Bonus: Conspiracy Theory

Conspiracy Theory: Docker Moby drastically and embraced. #Docker #DevOps

- Chris Short (@ChrisShort) December 29, 2017

I tweeted a theory about all the suspicious incidents with Docker in 2017: maybe Docker is aware that the end of the company is near. At a time when judging by organizational changes, the company is preparing to cease to exist (and, most likely, will be bought by someone big), the technical core of the company also decides to reorganize. The transfer of ontainerd CNCF , Moby as an upstream for Docker and a sudden friendship with Kubernetes - looks like an attempt to save all the good things that were done by the guys from Docker. This will allow you to sell a company to a large organization, such as Oracle or Microsoft, and not worry that the technological achievements of Docker employees will be buried under licensing restrictions. It will be better for developers and for the company. Needless to say, 2018 will be a very interesting year for Docker.