"Container design patterns" don't depend on containers

Lets make a kubernetes in a kubernetes on a kubernetes 🤔

Well I was kinda getting at the idea that nothing in that paper depends on, needs, or is enhanced by containerization per se. They're just basically co-process design patterns, and I'm not even sure they need that.

That there ARE public containers that do them is kinda besides the point as they ultimately just devolve into one unix process talking to another unix process. If you could do them with containers, you could do them with unix processes. It's not the containerization that makes it convenient, but the environment setup / teardown (i.e. rapid cheap provisioning)

While there ARE advantages to containerized apps, it's just not in the container design patterns themselves. If anything, those appear to be workarounds to the limitations that containerized environments lend you to.