Many years ago when a claim was made by a leading hosting website that LiteMage performed much better than other top hosting apps‚ it was viewed suspiciously and many questioned the methodology and outcome.  A retest was done to check this benchmark aging using the latest version of each software system.

Testing On A Real Ecommerce Business Situation

Magento 1.9 running on LiteSpeed‚ Apache and Nginx was chosen to know which web server is best suited to deal with the challenges posed by today’s complex ecommerce business environment. The challenge was to test three server setups with and without a caching solution to simulate a situation of 50 users accessing a Magento store‚ all at the same time. Every user was made to access 10 random catalogue pages from the total list of items. This helped us create a real-world ecommerce business situation.

We Used the Following Systems to Carry Out the Benchmark Test.

  • LiteSpeed Web Server 5.0.13 with LiteMage Cache 1.0.16
  • Nginx 1.8.1 with Varnish Cache 4.0.3 + Turpentine 0.6.8
  • Apache 2.2.15 with Varnish Cache 4.0.3 + Turpentine 0.6.8
  • LiteSpeed Web Server 5.0.13
  • Nginx 1.8.1
  • Apache 2.2.15

The server configuration used was 2-core Intel-R Quad Core with 4G RAM and running on CentOS 6.7. The client system was Debian GNU/Linux 8 and had the configuration of 2-core Intel-R Quad Core with 4 GB RAM. The following procedure was followed.

How The Test Was Conducted

We installed the Magento extension and configured it and used a browser to check the working of the cache. The appropriate command was given to warm up the cache and a few URLs were randomly checked with the browser to ensure that the Cache is hit and the same is reported in the response. The next obvious step was to run the benchmark test to stimulate the ecommerce market condition of 50 clients accessing 10 cached pages simultaneously. The same test was run several times and the result from one that was closest to the average was used.

The results showed that LiteSpeed+ LiteMage was the clear winner and that too by a wide margin. The test also revealed one important piece of information. Caching can make a huge difference regardless of the web server being used. Maximum transaction time test also revealed LiteSpeed plus LiteMage did better than others with a speed of 0.87 seconds against the 9.82 seconds of the worst performer which was Apache. Also‚ the longest transaction time improved but by a narrow margin for every server when using cache. Web servers using Varnish cache improved marginally but LiteSpeed Web Server reduced its time by nearly 90 per cent of the longest transaction taken while using cache.

According to hosting experts‚ network bandwidth is not a deciding factor in any well-performing application. These throughput tests used a 10GB switch and 10GB cards on both the client and server machines. These results delivered were based on software limitations.

This is yet another example which clearly shows how caching can put your web server on an entirely different level. While LiteSpeed + LiteMage outperformed the competition repeatedly‚ the throughput for every server improved. In almost every case‚ the throughput improved by more than 5X when cache was used.

What The Results Conclude

It is clear from the test that LiteSpeed and LiteMage is the best performer among all server setups used to test how web server performs when serving random pages to 50 simultaneous users. Are you interested in running a Magento store, and want to run it with a caching solution, LiteMage cache is the best option anyway you look at it.

Just to rule out any possibility of LiteMage emerging the winner inadvertently‚ the test was extended to 100 page view. By using a different testing method and testing tool‚ are the results impacted differently? One piece of advice that experts offer is to never use ‘Apache Bench’ for this type of testing. Siege is capable of following cookies while Apache Bench cannot. You cannot stimulate traffic for 50 visitors with Apache Bench commands when every visitor is accessing 10 catalogue pages simultaneously.

LiteSpeed is acknowledged as a lightweight and high performance web server and is event driven. The open source variant called OpenLiteSpeed can be used for testing. Installation of the system is quite easy and there is a neat web interface to manage everything from. You can leverage all the top features of Magento hosting system with LiteSpeed web server.