Before You Read the Performance Test Reports...
This page exhibits the performance test results under various conditions (e.g. various protocols and system environments). Please contact us if you have any specific performance test results to publish for your MINA-based application.
Apache MINA 2.0.0-M1-SNAPSHOT + AsyncWeb 0.9.0-SNAPSHOT
- Tested keep-alive mode using ApacheBench.
- Content length: 128 (excluding the header)
- Pentium 4 3GHz
- Ubuntu Linux 6.10
- 2 dual-core Opterons (4 cores, 270 Italy)
- Gentoo Linux 2.6.18-r6 x86_64
- 100Mbit Ethernet (direct link)
- Sun Java HotSpot(TM) 64-Bit Server VM (build 1.6.0-b105, mixed mode)
-server -Xms512m -Xmx512m -Xss128k -XX:+AggressiveOpts -XX:+UseParallelGC -XX:+UseBiasedLocking -XX:NewSize=64m
To show the performance characteristics of Apache MINA doesn't differ with the production-ready Web servers, the same test has been run on the Apache HTTPD 2.0.58. Because I don't know how to write an Apache HTTPD module, I simply used a dummy static file. Because the amount of the response header two HTTP servers generate is different, I changed the AsyncWeb to generate more traffic in the content. The size of one response was about 405 bytes.
The client machine in my company doesn't have 1Gbps Ethernet adapter nor a gigabit-capable CPU, I was not able to increase the content size. I made sure the network didn't saturate while the test at least.