Docker does not use LXC as default execution environment anymore. They created their own, called libcontainer¹. But with the new opencontainers movement, the package has been moved to runc².
That's why I used the word took, Docker used LXC as a base till version 0.9, untill it got enough traction, at which point it basically recreated LXC with libcontainer.
But that was not the point. The point is you have always had perfectly usable end user containers from the LXC project even before Docker. Then a VC funded company Docker bases itself on LXC and markets itself aggressively and suddenly a lot of users think LXC is 'low level' or 'difficult to use'? This messaging is coming from the Docker ecosystem and the result is the user confusion we see on most container threads here.
Informed discussion means people know what OS containers are, what value they deliver, and what Docker adds on top of OS containers so there is less confusion and FUD, and users can make informed decisions without a monoculture being pushed by aggressive marketing.
But that discussion cannot happen if you are in a hurry to 'own' the container story and cannot acknowledge clearly alternatives exists and what value you are adding exactly on top. I see people struggling with single app containers, layers and lack of storage persistence when they are simply looking to run a container as a lightweight VM.
The 'open container movement' is yet one more attempt to 'own' the container technology and perpetuate the conflation of Docker to containers. How can a 'open container movement' exclude the LXC project that is responsible for the development of much of the container technology available today. It should ideally be called 'Open App Container' because there is a huge difference between app containers and OS containers. OS containers provide capabilities and deployment flexibility that app containers simply cannot give because they are a restrictive use case of OS containers. Containers technology as a whole cannot be reduced to a single PAAS centric use case.
But that was not the point. The point is you have always had perfectly usable end user containers from the LXC project even before Docker. Then a VC funded company Docker bases itself on LXC and markets itself aggressively and suddenly a lot of users think LXC is 'low level' or 'difficult to use'? This messaging is coming from the Docker ecosystem and the result is the user confusion we see on most container threads here.
Informed discussion means people know what OS containers are, what value they deliver, and what Docker adds on top of OS containers so there is less confusion and FUD, and users can make informed decisions without a monoculture being pushed by aggressive marketing.
But that discussion cannot happen if you are in a hurry to 'own' the container story and cannot acknowledge clearly alternatives exists and what value you are adding exactly on top. I see people struggling with single app containers, layers and lack of storage persistence when they are simply looking to run a container as a lightweight VM.
The 'open container movement' is yet one more attempt to 'own' the container technology and perpetuate the conflation of Docker to containers. How can a 'open container movement' exclude the LXC project that is responsible for the development of much of the container technology available today. It should ideally be called 'Open App Container' because there is a huge difference between app containers and OS containers. OS containers provide capabilities and deployment flexibility that app containers simply cannot give because they are a restrictive use case of OS containers. Containers technology as a whole cannot be reduced to a single PAAS centric use case.