Overview of In-Memory Caching

Today’s Internet users are accustomed to instant page loads, even on their mobile devices. Databases running from disk are well suited for most use cases but realistically, all servers are limited in their processing power and can only handle a finite number of queries before scaling is required to allow the database engine to scale with the massive amounts of data.

Two simple rules apply to most web systems and server setups:

1. Certain objects are accessed more frequently than others and the database engine spends most of its resources on the most accessed objects.
   For example, Justin Beiber’s Tweets are accessed much more frequently than those from the average Twitter user. Certain Wikipedia articles are read considerably more than others.
2. RAM is faster for reading and writing data than SSD or HDD drives
   Modern RAM can reach read speeds of 5-7 GB/ s while the fastest SSD drives max out at 200-300 mb/ s. In addition, data is typically read from disk by the database engine which incurs a resource overhead when retrieving data.
   

In-Memory Caches capitalize on the these profound realizations by caching the most commonly accessed database data and page content in RAM so that users load the most used objects (almost) instantly and the database is freed from the constant, repetitive tasks of looking up the same values over and over again.