Designing Distributed Systems, 2nd Edition

Chapter 2. Important Distributed System Concepts

Before we get too deep into the design of distributed systems, it is valuable to ground ourselves in some key concepts that underpin the development of these distributed systems. If you are already experienced in distributed systems or systems design in general, some of these ideas may be familiar to you, and you may want to skip ahead, but hopefully at the completion of this chapter, you will have a core foundation for the patterns and designs presented throughout the rest of the book.

APIs and RPCs

Application programming interfaces, or APIs, form the core of any distributed system. You may have seen APIs in the context of programming language libraries or calls to operating system APIs. In distributed systems the idea is similar, but the API calls take place over the network between remote services. These calls are referred to as “RPCs” or “remote procedure calls.” RPCs rely on an underlying network transport, commonly, though not exclusively, the HTTP protocol and the JSON object format. Numerous other more structured protocols have existed over the years, from CORBA to gRPC as well as database-specific protocols. Though such systems provide capabilities beyond HTTP + JSON (also referred to as REST or RESTful), these continue to be the most popular because of their wide support, simplicity, and ease of integration with web-based frontends.

APIs can be synchronous or asynchronous. Synchronous APIs return only after the request ...