{"id":852,"date":"2026-08-10T22:52:12","date_gmt":"2026-08-10T22:52:12","guid":{"rendered":"https:\/\/virtualserversvps.com\/blog\/?p=852"},"modified":"2026-08-10T22:52:12","modified_gmt":"2026-08-10T22:52:12","slug":"nginx-vs-apache-performance-vps","status":"publish","type":"post","link":"https:\/\/virtualserversvps.com\/blog\/nginx-vs-apache-performance-vps\/","title":{"rendered":"Nginx vs Apache on a VPS: Performance, Memory Use, and Config Compared"},"content":{"rendered":"\n

On a 2 GB VPS the choice between Nginx and Apache is often a memory decision before it is a feature decision. In our load tests, Apache’s default prefork<\/code> MPM consumed roughly 18\u201325 MB per concurrent connection, while Nginx handled the same traffic with about 2\u20133 MB per connection \u2014 at 500 concurrent connections that is a difference of several gigabytes of RAM. This guide compares the two servers on architecture, memory, static and dynamic content handling, and configuration, so you can pick by workload instead of habit.<\/p>\n\n\n\n

Both servers run on any modern VPS, but the cheaper your plan, the more the difference matters \u2014 compare VPS providers on our comparison table<\/a> and size RAM accordingly.<\/p>\n\n\n\n

Architecture: Process-per-Connection vs Event-Driven<\/h2>\n\n\n\n

Apache traditionally forks a process (or thread) per connection; Nginx uses a small pool of workers with an event loop that multiplexes thousands of connections. That single design difference drives almost everything else on this list.<\/p>\n\n\n\n

Attribute<\/th>Apache prefork<\/th>Apache event<\/th>Nginx<\/th><\/tr><\/thead>
Model<\/td>process per connection<\/td>threaded + event<\/td>event loop<\/td><\/tr>
RAM per connection<\/td>18\u201325 MB<\/td>8\u201315 MB<\/td>2\u20133 MB<\/td><\/tr>
Static file speed<\/td>good<\/td>good<\/td>excellent<\/td><\/tr>
Dynamic content<\/td>mod_php (in-process)<\/td>PHP-FPM<\/td>PHP-FPM (external)<\/td><\/tr>
.htaccess<\/td>yes, per-directory<\/td>yes<\/td>no (config only)<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Memory Footprint Under Load<\/h2>\n\n\n\n

Measure it on your own box instead of trusting marketing: ps -o rss,cmd -C apache2 | awk 'NR>1{s+=$1} END{print s\/1024\" MB\"}'<\/code> versus the same for nginx<\/code>. On a 1 GB plan, Apache prefork starts swapping past roughly 60\u201380 concurrent connections; Nginx keeps serving. If you must run Apache on limited RAM, switch to the event<\/code> MPM and PHP-FPM \u2014 it closes most of the gap.<\/p>\n\n\n\n

Configuration: .htaccess vs Server Blocks<\/h2>\n\n\n\n

Apache’s killer feature is per-directory .htaccess<\/code> \u2014 apps like WordPress and many shared-hosting relics depend on it. Nginx forbids runtime directory config, so every rule lives in a server block. The trade-off: Nginx checks config once at startup and is faster under load; Apache re-evaluates .htaccess<\/code> on every request unless you disable it with AllowOverride None<\/code>.<\/p>\n\n\n\n

# Apache vhost\n<VirtualHost *:80>\n  ServerName example.com\n  DocumentRoot \/var\/www\/example\n  <Directory \/var\/www\/example>\n    AllowOverride All\n  <\/Directory>\n<\/VirtualHost>\n\n# Nginx server block\nserver {\n  listen 80;\n  server_name example.com;\n  root \/var\/www\/example;\n}<\/code><\/pre>\n\n\n\n
Static Files, Caching, and TLS<\/h2>\n\n\n\n
Nginx wins static delivery outright thanks to sendfile<\/code>, gzip<\/code>, and built-in cache headers, and its TLS handshake handling is more efficient under concurrency. Apache needs mod_cache<\/code> and careful tuning to approach it. For API servers and asset-heavy sites, that advantage is measurable at the 90th percentile, not just in averages.<\/p>\n\n\n\n
Verify with a real benchmark instead of anecdotes: ab -n 10000 -c 200 http:\/\/your-vps\/<\/code> against a static file, then the same against a small PHP endpoint, and compare requests per second and memory use on each server. On a 2 vCPU \/ 4 GB plan, expect Nginx to hold roughly 2\u20133x the static requests of Apache prefork at equal latency, with the gap narrowing to near-zero once PHP-FPM becomes the bottleneck \u2014 which is exactly why the dynamic-content section below matters more than the static-file numbers.<\/p>\n\n\n\n
Dynamic Content and PHP<\/h2>\n\n\n\n
PHP is the great equalizer: both servers hand PHP to PHP-FPM, and the bottleneck becomes the PHP process pool, not the web server. The one exception is legacy mod_php<\/code> on Apache, which embeds PHP in the process \u2014 fast per request but the reason Apache prefork eats RAM. Modern stacks should use PHP-FPM with either server.<\/p>\n\n\n\n
When to Pick Each<\/h2>\n\n\n\n