15.1. NoSQL Database Overview

The term NoSQL encompasses a broad range of databases, each with its own strengths and weaknesses, and most with wildly different goals and use cases. When looking at the NoSQL ecosystem today, the databases tend to break down into five broad categories: pure key/value, data structure, graph, document oriented, and highly distributed. Each of these categories of databases has a slightly different use case, and each category makes different trade-offs. We're going to go through them and take a look at those trade-offs.

15.1.1. Pure Key/Value

Pure key/value databases have actually been around for a very long time. Even before SQL databases had risen in popularity, dbm (a pure key/value database) was in use on many Unix systems around the world. After that, there was Berkeley DB, which is still a maintained and viable database solution. Today these pure key/value stores are seeing a resurgence in popularity, thanks in part to the resurgence in popularity of all NoSQL databases but also because of the development of some faster and more modern implementations such as Tokyo Cabinet, Kyoto Cabinet, and MemcacheDB.