Understanding Apache Work Modes and Nginx Advantages for High‑Concurrency Web Servers
The article explains Apache's three MPMs (prefork, worker, event), how Nginx's event‑driven architecture and epoll model improve concurrency handling, and provides performance comparisons showing Nginx handling tens of thousands of connections with far lower memory usage than Apache.
Top architect explains Apache's three MPMs (prefork, worker, event) and their working principles.
Prefork uses multiple processes, each handling one request; worker uses multithreading; event uses asynchronous I/O with epoll.
To improve web server concurrency, conditions include thread‑based processing, event‑driven model, AIO, and mmap support, all of which Nginx implements.
Nginx's architecture includes a master process and several worker processes (plus optional cache loader/manager), each single‑threaded and communicating via shared memory.
Benchmarks show Nginx + PHP(FastCGI) handling over 30,000 concurrent connections with modest memory usage (e.g., 15M*10=150M for workers, 20M*64=1280M for php‑cgi), far outperforming Apache prefork which exhausts memory at ~3,000 connections.
The article also compares I/O multiplexing models: select (limited FD count), poll (no limit), and epoll (event‑driven, efficient, Linux‑only), explaining why Nginx’s epoll‑based design solves the C10K problem.
Overall, Nginx’s event‑driven, non‑blocking, single‑threaded workers provide higher performance and lower resource consumption than Apache’s process‑ or thread‑based models.
Top Architect
Top Architect focuses on sharing practical architecture knowledge, covering enterprise, system, website, large‑scale distributed, and high‑availability architectures, plus architecture adjustments using internet technologies. We welcome idea‑driven, sharing‑oriented architects to exchange and learn together.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.