LiteSpeed Web Server

LiteSpeed Web Server (often shortened to LiteSpeed or LSWS) is a commercial, event-driven HTTP server developed by LiteSpeed Technologies that is designed as a drop-in replacement for Apache: it reads Apache configuration files and .htaccess rules while serving requests with a far smaller memory and process footprint. Among web servers, it is especially common in shared and reseller hosting, where compatibility with existing Apache setups matters as much as raw performance.

Its distinguishing features are a built-in server-level page cache called LSCache, a dedicated PHP interface called LSAPI, and early support for HTTP/3 over QUIC. A free, open-source sibling, OpenLiteSpeed, shares the core engine but not the Apache compatibility layer.

How LiteSpeed handles requests

LiteSpeed is built around an event-driven core. Instead of dedicating a process or thread to every client, as the prefork model of Apache does, it runs a few workers that multiplex very large numbers of sockets with non-blocking calls. When many visitors connect at once, or connect over slow mobile links, memory consumption stays nearly constant, and the server does not run out of processes the way a thread-per-connection design can.

Static files are served directly by the workers. Dynamic requests go to external application servers: PHP through LSAPI, and other languages through FastCGI, proxying or language-specific interfaces. Because LiteSpeed understands Apache’s mod_rewrite syntax, existing rewrite rules in .htaccess files keep working, and the server caches parsed .htaccess content so it does not reread the files on every request.

Configuration can follow two paths. In Apache compatibility mode, LSWS reads httpd.conf and virtual host files generated by control panels such as cPanel, Plesk or DirectAdmin. In native mode, administrators configure it through its own XML files or the WebAdmin console, which listens on port 7080 by default.

PHP handling with LSAPI

LSAPI is LiteSpeed’s own application interface, used mainly with a PHP build called lsphp. It is comparable in purpose to FastCGI with PHP-FPM, but it is integrated more tightly with the server’s process management and resource limits, which suits hosting environments with many separate accounts.

In shared hosting, LSAPI is often combined with CloudLinux resource limits and suEXEC-style user switching, so each account’s PHP processes run under its own system user and cannot exhaust the whole server.

LSAPI can also run PHP in different process modes. Worker mode keeps a pool of PHP processes ready, ProcessGroup mode starts a parent process per account whose children share opcode cache, and daemon mode forks requests from a persistent parent. ProcessGroup mode is popular on shared servers because it keeps the opcode cache warm per account while preserving isolation between users.

LSCache and page caching

LSCache is a full-page cache built into the server. Instead of passing every request to PHP, LiteSpeed can serve a stored copy of a rendered page directly, which reduces time to first byte and CPU use dramatically for pages that do not change per visitor.

The cache relies on application plugins, such as LSCache for WordPress, to tell the server what to cache, for how long and when to purge. Plugins tag pages so that, for example, publishing a post purges only the affected pages. They also support private caching for logged-in users and edge-side includes for personalized fragments, plus optional image optimization and CSS or JavaScript minification.

The key limitation is that LSCache works only on LiteSpeed servers. A site that moves to Nginx or Apache must switch to another caching method, and running two page caches at once, for example LSCache plus another caching plugin, often causes stale or broken pages.

LiteSpeed compared with other web servers

LiteSpeed competes with both Apache, whose configuration it imitates, and Nginx, whose event-driven design it shares. The table compares them with OpenLiteSpeed.

Performance comparisons depend heavily on configuration and workload. LiteSpeed’s advantage is most visible on PHP-heavy sites with caching enabled and on servers hosting many Apache-configured accounts, where replacing Apache requires no rewrite of configuration.

Choosing and running LiteSpeed

LiteSpeed is licensed by number of domains, CPU cores or memory, depending on edition, with a free tier limited to a small single site. The table summarizes when it fits.

Common pitfalls include unsupported Apache modules, which LiteSpeed silently ignores or approximates, license limits that cap worker processes on busy servers, and cache plugins left with default rules that cache pages containing user-specific content. Testing critical rewrite rules and cache exclusions after migration prevents most problems.

In plain words

A web server is the program that sends web pages to visitors. LiteSpeed is one such program. It can take the place of Apache, the most traditional choice, without changing how websites are set up.

Its strength is speed under load. It keeps ready-made copies of pages, so it does not rebuild the same page for every visitor. Think of a kitchen that keeps popular dishes ready instead of cooking each order from scratch.

A real-world example

A hosting provider runs about 400 small business websites, mostly WordPress, on a cPanel server with Apache. During seasonal traffic peaks, the server often hits memory limits and slow page loads. The provider installs LiteSpeed Web Server, which reads the existing Apache configuration, so no account needs changes.

Customers enable the LSCache plugin, and cached pages are now served without starting PHP. In monitoring, average time to first byte drops from roughly 800 ms to under 150 ms for cached pages, and peak memory use falls enough that the provider postpones a planned hardware upgrade. Existing .htaccess redirects keep working without edits.