prefork

Prefork is a server architecture used in some web servers where the master process spawns a pool of child processes. Each child handles a single client connection at a time, and the master forks new children as demand increases, up to configured limits. This approach emphasizes process isolation and simplicity.

Historically, prefork is associated with the prefork Multi-Processing Module (MPM) of the Apache HTTP Server. In this model, threading is avoided entirely, with requests served by separate processes rather than by threads within a single process.

Advantages of prefork include compatibility with non-thread-safe modules, such as older versions of PHP and certain database drivers. The isolation between processes reduces the risk that a failure in one request affects others, and memory protection is straightforward because each process has its own address space.

Disadvantages include higher memory usage, since each process loads its own copy of modules and data. Concurrency is limited by the number of processes, and context switching can become a bottleneck under heavy load. This model is generally less efficient on modern hardware for high traffic due to its inability to leverage threads for sharing resources.

In contemporary deployments, many servers favor threaded or event-driven models (such as Apache’s worker or event MPMs or Nginx) to improve scalability and resource utilization. Prefork remains in use when stability and compatibility with legacy modules are prioritized.

See also: Apache HTTP Server, Multi-Processing Module, mod_php, Worker MPM, Event MPM.