Java is a very common platform choice for running production code for many applications across many domains. When data scientists create a model for existing applications, Java is a natural choice, since it can be seamlessly integrated into the code. This case is straightforward, you create a separate package, implement your models there, and make sure other packages use it. Another possible option is packaging the code into a separate JAR file, and include it as a Maven dependency.
But there is a different architectural approach for combining multiple components of a large system--the microservices architecture. The main idea is that a system should be composed of small independent units with their own lifecycle--their development, testing, and deployment cycles are independent of all other components.
These microservices typically communicate via REST API, which is based on HTTP. It is based on four HTTP methods--GET, POST, PUT and DELETE. The first two are most commonly used...