“The Docker fork talk is a storm in a teacup”
Discussions about a possible Docker fork have taken the industry by storm. We have the group which supports Docker 1.12 and points out its strengths and the group which claims that what the world needs right now is “boring core infrastructure.” We asked Benjamin Wootton, the co-founder of Contino, to tell us what he thinks about Docker 1.12 and to weigh in on this Docker fork discussion.
The discussions about a possible Docker fork are gaining momentum. People are stating their views on this issue via Twitter, Medium, personal blogs,forums and more and pro- and anti-Docker fork groups are being formed.
We invited Benjamin Wootton, the co-founder of Contino, to weigh in on the rumors and share his experience with Docker 1.12.
JAXenter: Have you tried Docker 1.12? What was your first impression of it?
Benjamin Wootton: Our first impression of 1.12 was that the headline feature of 1.12, integrating Swarm into Docker was a big success. Setting up orchestration frameworks can be pretty full on, particularly if you want to do it in a secure way. SwarmKit makes it very simple. In one command you can stand up a master, and in another command you can add a node. There is no separate software, no key value store, and it’s all secured under the hood. Our first impression is that this would really drive adoption of Swarm and container orchestration.
JAXenter: There is a Hacker News thread about Docker 1.12’s flaws. Did you find anything wrong in it?
Benjamin Wootton: We think this is a bit of a storm in a teacup. Most of the blog posts are referring to stability in Swarm rather than Docker, but these are generally edge cases such as running on a Raspberry Pi. SwarmKit was a big, ambitious project and a big merge, and though we all want perfect bug-free software first time, I think we can be pragmatic in expecting a few bugs on edge cases for a first release of such a widely watched project.
I would also point out that the integrated Swarm in 1.12 is still an optional feature and at the current time there is no plan to deprecate standalone Swarm, so the old way still works fine.
SEE ALSO: Can Docker be ousted?
JAXenter: There is a relatively large discussion going on about a possible Docker fork. What’s your take on that? Do we need boring code infrastructure or should Docker continue to innovate and grow?
Benjamin Wootton: We also don’t think this has any legs — it was the one New Stack article which was full of speculation.
Our experience since deploying Docker into production around the 0.8 timeframe is that the container building block has been a very stable platform to build upon. SwarmKit is the biggest change in a while, but even that is completely optional and doesn’t impact building or running individual containers.
I also don’t think a fork is viable because Docker is delivering and innovating so quickly. Any fork or standard would fall behind pretty quickly and would fail to get momentum in the industry.
Thank you very much!