This benchmark compares the speed at which different web servers respond to requests for small PHP scripts using both non-keep-alive and keep-alive connections.
Summary
| Concurrent Users | Keep-Alive | Nginx / FPM | Apache / FPM | Apache / EA4 Default |
|---|---|---|---|---|
| 10 | No | 1.4x | 3.4x | 127.4x |
| Yes | 1.5x | 3.6x | 192.6x | |
| 100 | No | 1.4x | 3.4x | 137.5x |
| Yes | 1.5x | 4.1x | 197.5x |
Notes:
- We used a simple PHP hello world script, 4.0K in size. We used such a tiny script to avoid saturating the network connection and to show raw speed differences between the different setups.
- Part of the difference in speeds is due to different server APIs. LiteSpeed Enterprise used LSAPI, PHP-FPM uses FCGI, EA4 Default uses CGI.
- Opcode caching was used for all setups.
- The benchmark simulated serving 10000 requests to 10 and 100 users.
- The test was performed over a 10GBps network connection to make sure network bandwidth did not become a bottleneck.
Test Environment
Software:
LiteSpeed Enterprise 5.1.13 Nginx 1.10.3 Apache 2.4.25 PHP 7.1.2 cPanel 62
Server hardware specs:
Intel Xeon X5660 Single Core @ 2.80GHz 2GB RAM 50GB HDD CloudLinux 7.3
Client hardware specs:
Intel Xeon E5-1620 Dual Core @ 3.60GHz 2GB RAM CentOS 6.8 with OpenVZ kernel 2.6.32-042stab116.1 Intel X540 10GBASE-T on board NIC
Network Switch:
Netgear XS708E-100NES 8-ports 10G switch
We welcome your feedback on our forum.
