Optimizing For Production Workloads

Breaking down the containers runtimes into their base functionality and then building them up into a series or core libraries and tools to specialize in core capabilities. Our goal is, rather then have one monolithic daemon to do all container management, to build up a series of tools that specialize in each different function of container management.

The tooling landscape for containers is evolving rapidly--to keep things running smoothly in production, you need to keep your eye on the latest developments in container infrastructure. This talk will introduce the following:

1. containers/Image: Library used by other tools for pulling and pushing container images to and from a container registry like docker.io and into other types of container storage
2. containers/storage: Library used for storing container images on disk supporting multiple container backends including Overlay, Devicemapper, and Btrfs
3. libpod : Library for managing the lifecycle of pods and containers
4. Skopeo: CLI tool that uses containers/image and containers/storage to move container images between different types of containers storage, including container registries, container storage, Docker storage, OCI Directories etc. Think of it as scp for container images.
5. Buildah: CLI tool that allows you to build OCI container images using standard linux tools, no daemon necessary, also has support for Dockerfile
6. CRI-O : API server daemon for the Kubernetes Container Runtime Interface. optimized for serving the Kubernetes and OpenShift Workloads. Supports multiple container runtimes like runc and kata containers out of the box.
7. Podman: Simple CLI tool used to create pods and containers without requiring a daemon.

Tools share the same container storage and infrastructure, allowing users to pick and choose the best tool for their workloads.