RESTful Web API Patterns and Practices Cookbook

Chapter 6. Distributed Data

First step in breaking the data centric habit, is to stop designing systems as a collection of data services, and instead design for business capabilities.

Irakli Nadareishvili, JPMorgan Chase

This chapter is devoted to recipes for data-centric service interfaces. Data-centric interfaces need to follow all the same principles covered in Chapter 5 along with some additional details that come from the responsibility of storing and managing data. These details involve assuring data integrity, hiding internal data models and implementation technology, and dealing with a wide range of possible network failures without invalidating any existing data. This is especially important now that data can easily travel around the world and come into contact with regulations such as General Data Protection Regulation (GDPR) and others.

Not all services need to manage their own data, but most have some level of data support responsibilities. The challenge of data-centric services is that they typically support persistent data. Even when a service goes offline, the data must continue to exist and/or be accessible by other services. In some cases the service interface has the task of mixing locally managed data with data from other external services. This compounds the integrity and reliability problem since the target interface must now rely on other data services to complete the request work. And, especially in the case of writing data, the more services involved ...