NGINX Architecture

NGINX is an open source reverse proxy and web server designed for scale. It exploded in popularity as the first line of defense to backend infrastructures. Whether as a caching layer, load balancer, an API gateway or Web server, NGINX does it all.

In this article I explore the internal architecture of NGINX focusing on the following topics:

• NGINX worker processes
• Connection Management
• Request Processing

Worker Processes

The major part of NGINX architecture is the master process. When you spin up NGINX, it starts a process that manages the rest of processes in NGINX.

The master process creates two processes for cache management. One for reading the cache from disk and another for refreshing the caches. What interests me the most are the worker processes. These are the power horses that do most of the work in NGINX.

When you set NGINX to auto mode (the default), the application spins up a number of worker processes based on how many CPU cores you have. This changes based on whether you have Hyper-Threading enabled or not.