What Are Miniservices?

Modern computing can be described as the compromises between processing (in-process execution), communication (between networked processes), and data (input/output of a process).

There are many other ways of defining and breaking out computing, but I have found these three aspects of a computing system pivotal to design and architecture discussions, so I will be highlighting them throughout this report.

For each of these categories, there are critical decisions to make within an architecture. All three are similarly important, yet I have consistently seen that teams attempting to build a microservices architecture tend to focus primarily on processing and communication. Conway’s Law, which observes that organizations build systems that resemble how they communicate, in practice often prevents a good discussion about the data. You can find effective strategies, books, and tooling on the fallacies of distributed computing (communication) or on building and deploying fine-grained interfaces on atomic services (processing), but the data topic too often remains siloed and disconnected from the conversation.

Often, I see data design relegated to a few “data people” who are usually on an external/shared team. As a result of this artificial distinction, discussions about the data life cycle are a hodgepodge of conversations specific to one technology or tool set, limited to a specific pattern of data access, and isolated from the wider architectural conversation ...