Most developers use Docker. Few actually understand how it works under the hood. Here is the simple flow 👇 > You send commands like docker pull or docker run. > These go to the Docker Daemon via REST API > The daemon does the heavy lifting building, running, managing containers > Images are pulled from (or pushed to) a registry like Docker Hub That is it. Client → Daemon → Registry. #Docker #DevOps #Backend #SoftwareEngineering #FullStack
Also docker build does package the full build context (all field added and copied during build) and sends it to the demon, so sou don't even need a demon running on the same machine you run docker build
This is a great overview!
Simple and accurate. The daemon being a separate process is also what explains why you need sudo or the docker group on Linux — you're talking to a background service over a socket, not running Docker directly.
Check out WolfStack - it can run Docker/kubernetes and has a really neat network interface to configure everything.
While that’s an accurate high level overview. The devil is in the lower level - how different platforms bridge between the OS and container and how that impacts performance. If you really want a deep dive, take a look at what the Amazon Web Services (AWS) engineering team is doing with fargate. This article is few yrs now, but still a good overview of the guts of how containers actually work and how they can be tuned for performance: https://aws.amazon.com/blogs/containers/under-the-hood-fargate-data-plane/