The latest container orchestrators have finally driven us to think for abstraction without worrying about the underlying concrete implementation: until we're talking about containers we know the 'black magic' behind the hood (namespaces, cgroups, overlayfs). Even the container approach helped the world to leverage applications delivery it has become obsolete (sic) since the newest cloud-native architectures require complex patterns such as Ambassador, Configurator, Adapter and much more. People assume that a Pod is just a container, often: obviously this is not true... and by the way, who knows what is a Pod exactly, how is made of, and how Kubernetes orchestrates it? The goal of this talk is to provide a low-level description of what a Pod is and how Kubernetes help us to work with this kind of resource in order to approach container patterns, demonstrating with a live coding demo how to build a simple Pod in order to understand the Kubernetes internals that provides this computational resources abstractions.

свернуть