From the abstract:
A fantastic, fictional, fable of a distributed system and the tooling and teams who herd it. Follow two teams learning that the design of their systems includes everything about how it is developed, tested, deployed, and managed — not just what it looks like in production.
Along the way, learn about the tools and practices that enable their eventual success — in particular Atlas, a new, open source, systems modeling and management tool developed by Ning.
This talk is a deep dive into the ideas behind Atlas, as well as the particular implementation, which embraces the fact that parts of real systems are always being built at different speeds, by different people, and run in many different places at the same time.