Learning Node

Chapter 9. Structured Data with Node and Redis

When it comes to data, there’s relational databases and Everything Else, otherwise known as NoSQL. In the NoSQL category, a type of structured data is based on key/value pairs, typically stored in memory for extremely fast access. The three most popular in-memory key/value stores are Memcached, Cassandra, and Redis. Happily for Node developers, there is Node support for all three stores.

Memcached is primarily used as a way of caching data queries for quick access in memory. It’s also quite good with distributed computing, but has limited support for more complex data. It’s useful for applications that do a lot of queries, but less so for applications doing a lot of data writing and reading. Redis is the superior data store for the latter type of application. In addition, Redis can be persisted, and it provides more flexibility than Memcached—especially in its support for different types of data. However, unlike Memcached, Redis works only on a single machine.

The same factors also come into play when comparing Redis and Cassandra. Like Memcached, Cassandra has support for clusters. However, also like Memcached, it has limited data structure support. It’s good for ad hoc queries—a use that does not favor Redis. However, Redis is simple to use, uncomplicated, and typically faster than Cassandra. For these reasons, and others, Redis has gained a greater following among Node developers, which is why I picked it over Memcached and Cassandra ...