These notes on Redis are based on its beta version. One could see its source code on my GitHub, where I have tidied the original codes in order to make the codes more easily readable with only around 3,000 code lines. 1. Key Data Structure
These notes on Redis are based on its beta version. One could see its source code on my GitHub, where I have tidied the original codes in order to make the codes more easily readable with only around 3,000 code lines.
1. Key Data Structure
Redis is a kind of key value store. So its key data structure is hash resident in main memory. Keys and values in Redis' hash would be serialized into disk when save command is executed. Here values could be strings, lists, and even sets (not implemented in the beta version). The following graph is borrowed from 《Redis 设计与实现》 to show Redis' key data structure, that is hash.
2. Event-driven Implementation
When a client connects to Redis, how does Redis know the client"s connection? How does Redis implement some periodic checks, such as saving database after certain number of changes or time, and resizing the hash? In essence, Redis runs in a while loop, where API select() is used to respond to the above situations.
