I got that reaction too when I read it. There were other, sneakier NLP-like use of language. For example, "Being at the forefront of the container wave, one trend we see emerging in 2017 is containers going mainstream ..." It's a way to sneak in the idea that Docker is "at the forefront of the container wave". By accepting the rest of the sentence, it is easy to accept the earlier clause without challenge.
If anything happened in 2016 in the containerization space, it was that Docker was playing a lot of catch-up to Kubernetes. It seems like they rushed out things and then they botched the OSX native release of Docker in Q2 2016, and lost a lot of goodwill in the community of individual developers (more so than CoreOS did with rkt? That's how it feels like to me. Back then, people thought CoreOS were the assholes for forking. I think that sentiment has shifted). I have not tracked how Docker is doing in enterprise, but I guess they are doing well?
I kinda doubt Docker (specifically, Docker Swarm, and the orchestration tools Docker forced into docker-engine) will ever catch up to Kubernetes, but hey, things move fast, right?
Having said all of that, I like this description: https://news.ycombinator.com/item?id=14141832 ... which makes _much_ more sense. Maybe now, someone will use Moby to create a native OSX Docker kit that actually has decent file sync performance (either with unisonfs, rsync, or nfs, and not try to layer it as another FS).
I agree, Docker is in a real pickle. Kubernetes has the momentum and it is decoupling from Docker rapidly.
Docker's biggest asset right now is that everyone has written Dockerfiles already, but unfortunately, Dockerfiles are terrible and people can't wait to get off of them.
Long-term, I expect the platform to be k8s + CoreOS + rkt. Docker's future, if anything, will be in hosting the image repository.
I have been using the Google container registry to store my private images. Costs just a few cents a month for my use case. If you don't want to pay the minimum 7$ per month for more than 1 private image.
> Long-term, I expect the platform to be k8s + CoreOS + rkt.
I have been dabbling with that same thought for a while too. One thing I have not yet figured out, though, is what CoreOS' vision is for using rkt as a developer on Mac, something a lot of developers are.
If anything happened in 2016 in the containerization space, it was that Docker was playing a lot of catch-up to Kubernetes. It seems like they rushed out things and then they botched the OSX native release of Docker in Q2 2016, and lost a lot of goodwill in the community of individual developers (more so than CoreOS did with rkt? That's how it feels like to me. Back then, people thought CoreOS were the assholes for forking. I think that sentiment has shifted). I have not tracked how Docker is doing in enterprise, but I guess they are doing well?
I kinda doubt Docker (specifically, Docker Swarm, and the orchestration tools Docker forced into docker-engine) will ever catch up to Kubernetes, but hey, things move fast, right?
Having said all of that, I like this description: https://news.ycombinator.com/item?id=14141832 ... which makes _much_ more sense. Maybe now, someone will use Moby to create a native OSX Docker kit that actually has decent file sync performance (either with unisonfs, rsync, or nfs, and not try to layer it as another FS).