CockroachDB: The Definitive Guide

Chapter 2. CockroachDB Architecture

The architecture of a software system defines the high-level design decisions that enable the goals of that system. As you may recall from Chapter 1, the goals of CockroachDB are to provide a scalable, highly available, highly performant, strongly consistent, geo-distributed, SQL-powered relational database system capable of running across a wide variety of hardware platforms. The architecture of CockroachDB is aligned to those objectives.

Feel Free to Skip Ahead!

The CockroachDB architecture is sophisticated: it incorporates decades of database engineering best practice designs together with several unique innovations. However, CockroachDB doesn’t require that you understand its internals to get things done. If you’re in a hurry to get started with CockroachDB, you can skip forward to the next chapter and return to this chapter later as necessary. We will, however, assume you are broadly familiar with the key concepts in this chapter when we consider advanced topics later in the book. Those key concepts are summarized over the next few pages and elaborated on in the remainder of the chapter.

There are multiple ways of looking at the CockroachDB architecture. At the cluster level, a CockroachDB deployment consists of one or more shared-nothing, leaderless nodes that collaborate to present a single logical view of the distributed database system. Within each node, we can observe the CockroachDB architecture as a series of layers that provide ...