Infrastructure as Code, 2nd Edition

Chapter 18. Organizing Infrastructure Code

An infrastructure codebase may include various types of code, including stack definitions, server configurations, modules, libraries, tests, configuration, and utilities.

How should you organize this code across and within projects? How should you organize projects across repositories? Do infrastructure and application code belong together, or should they be separated? How should you organize code for an estate with multiple parts?

Organizing Projects and Repositories

In this context, a project is a collection of code used to build a discrete component of the system. There is no hard rule on how much a single project or its component can include. “Patterns and Antipatterns for Structuring Stacks” describes different levels of scope for an infrastructure stack, for instance.

A project may depend on other projects in the codebase. Ideally, these dependencies and the boundaries between projects are well-defined, and clearly reflected in the way project code is organized.

Conway’s Law (see “Align Boundaries with Organizational Structures”) says that there is a direct relationship between the structure of the organization and the systems that it builds. Poor alignment of team structures and ownership of systems, and the code that defines those systems, creates friction and inefficiency.

The flip side of drawing boundaries between projects is integrating projects when there are dependencies between them, as described for stacks in Chapter 17 ...