There are many people advocating that we start building our applications as a monolith and then later extracting services out from the monolith with the argument being that we can focus on functional value as opposed to the tooling to manage and monitor a complex series of services.

However, what are the choices we can make when building the monolith that will allow us to subsequently extract microservices from the monolith with the minimum amount of effort?

In this talk, I'll explore patterns for refactoring/structuring the monolith to make it easier to extract into microservices, by building loosely-coupled message based systems within the monolith, and how to extract the components into microservices.