Shams Nahid
Shams Nahid's Blog

Shams Nahid's Blog

Redis 101: A super fast in-memory database

Redis 101: A super fast in-memory database

Cache data in a full-fledged in-memory database.

Shams Nahid's photo
Shams Nahid
·Nov 5, 2022·

2 min read

Table of contents

  • Why Redis
  • Design Methodology
  • Final Thoughts

Redis is one of the fasted in-memory databases out there and very much different from the traditional SQL/no-SQL databases. Initially, it was developed for caching, but the simplicity and speed take it to a completed in-memory database.

Why Redis

When we compare this Redis database with a traditional database, it is similar in a way,

  • Persist data for applications
  • We run queries and fetch data

But it does the job super fast of storing and retrieving data on the database by using the computer memory.

Redis is a super-fast temporary database, that uses computer memory, not the HDD or any other permanent storage hardware

Redis is fast, because,

  • All data is stored in memory
  • Use a very simple and straightforward data structure
  • Provides limited features for querying data

Redis narrowed down the data structures in it into LinkedList, hashMap, or set type structures.

So Redis is fast because it is simple.

Design Methodology

This simplicity brings a couple of challenges during we design a database,

  • Fitting data in a limited amount of memory
  • Make use of simple data structures to persist data
  • Utilize the database with its limited feature set

Redis has the following data sets,

  • String: Plain string or number
  • List: List of strings
  • Hash: Collections of key values
  • Set: Set of unique strings
  • Sorted Set: Set if sorted unique strings
  • Bitmap: Kind of collection of booleans
  • HyperLog: Kind of collection of booleans
  • JSON: Nested JSON structure
  • Index: Internal data, used for searching

It does also support concurrency with the watch and lock method.

For traditional databases, we design the database as follows,

  • First, we persist data in tables or collections
  • Determine what query we execute depending on what data we required

With Redis, we took the reverse approach,

  • First, figure out what queries we will perform
  • Design data according to that query

A couple of additional concerns with Redis are,

  • Types of data we are storing
  • Should we concern with data size (for example, static page/dynamic page)
  • When do we expire the data
  • How do we name the key
  • Any consideration of business logic

Final Thoughts

When it comes to a super fast in-memory database solution, Redis can be a no-brainer. Apart from Redis itself, all the major cloud service providers have managed Redis service.

 
Share this