01 Introduction

In many applications performance is an essential priority, and often a small delay in response time could mean a big financial loss.

For these kind of needs key-value stores were built, a type of db which guarantees the best possible lookup time of any storage system.

Key-value stores are build upon the assumption that any entity can be seen as a value pointed by a key, and so when you are looking for a certain value you need to search for the key associated to it.

In these contest keys are just used to retrieve a specific value, unlike relational databases where primary and foreign keys are means to do searches and mapping between tables; this obviously grants a significant speed boost during reads.

Key-Value stores can be used to improve performance at many different levels:

• Database
• Caching Layers
• Message Brokers

Caching

A cache is a collection of data duplicating original values stored elsewhere or computed earlier, where the original data is expensive to fetch (owing to longer access time) or to compute, compared to the cost of reading the cache.

Caches are only efficient when the benefits of faster access outweigh the overhead of checking and keeping your cache up to date.

There are only two hard things in Computer Science: cache invalidation and naming things. -- Phil Karlton.