There a multiple motivations behind distributing systems. Despite certain introduced challenges and overheads in communication, performance, and organization, distribution is often necessary.

In order to design the system landscape, the system's context map that represents the individual responsibilities needs to be taken into consideration. It's advisable to design application APIs in clear, lean interfaces, ideally implemented with standard communication protocols. Before introducing breaking changes, engineers as well as business experts need to ask themselves whether it is necessary to force client functionality to stop working. In the same way, APIs should be designed in a resilient way, preventing unnecessary breaks, in other words: be conservative in what you do and liberal in what you accept.

Engineers that build distributed applications need to be aware of the trade-off between consistency and scalability. The majority of applications that use synchronous communication involving an external system will likely scale well enough. Distributed transactions should be avoided.

In order to communicate asynchronously, application can be based on event-driven architectures. The CQRS principle combines the motivations behind event-driven architectures and event sourcing. Whereas CQRS certainly offers interesting solutions, it only makes sense if there is a need for distributing application.

Microservice architectures don't share common technology or data with each other. Shared-nothing architectures are free to choose implementations and persistence technology. Zero-dependency Java EE applications shipped in containers are a reasonable fit for microservices. The one application per application server approach matches the idea of shared-nothing architectures. There are many aspects in which Java EE applications running in container orchestration frameworks support developing microservice architectures, such as service discovery, resilient communication via timeout, circuit breakers or bulkheads.

The following chapter covers the topics of performance, monitoring and logging.