{"id":9518,"date":"2018-04-12T06:00:29","date_gmt":"2018-04-12T10:00:29","guid":{"rendered":"https:\/\/blog.litespeedtech.com\/?p=9518"},"modified":"2018-04-12T06:00:29","modified_gmt":"2018-04-12T10:00:29","slug":"benchmarks-litespeed-cache-module-for-drupal","status":"publish","type":"post","link":"https:\/\/lswp.store\/index.php\/2018\/04\/12\/benchmarks-litespeed-cache-module-for-drupal\/","title":{"rendered":"Benchmarking LSCache for Drupal"},"content":{"rendered":"

\"LiteSpeed<\/a><\/p>\n

We\u2019re back with more benchmarks! This time, we\u2019ve put our new LSCache module for Drupal 8<\/a> to the test. We compared LiteSpeed Web Server with LSCache<\/a>, Apache with Varnish (Cachewall), and Apache with System Cache.<\/p>\n

Our testing consisted of burst tests and load testing, both of which are explained in more detail below.<\/p>\n

The Results<\/h2>\n

Benchmark<\/h3>\n

The following chart shows how many requests per second each server + cache solution combination processed during our burst tests. The larger the number, the better.
\n\"\"
\nAs you can see, LiteSpeed Web Server + LSCache outperforms both Apache + Varnish and Apache + System Cache.<\/p>\n

Load Testing<\/h3>\n

The following graphs show the CPU and Memory usage during testing for each server + cache solution combination tested. For these measurements, a smaller number is better.
\n\"\"\"\"
\nLiteSpeed Web Server + LSCache consumed less memory and CPU than both Apache tests.<\/p>\n

Try it Yourself<\/h2>\n

You can see the specs of our configuration and test environment below. Feel free to run the same benchmarks yourself and verify our results.<\/p>\n

These are the commands we used in all tests (Replace example.com\/path<\/code> with the location of your test subjects):<\/p>\n

Benchmark<\/u><\/p>\n

ab -n 10000 -k -H \"Accept-Encoding: gzip,deflate\" -c 100 example.com\/path<\/pre>\n
Load Testing<\/u>
\nLSWS+LSCache and Apache+Varnish:<\/p>\n
siege -c 100 -r 20 -d 6 example.com\/path<\/pre>\n
Apache+System:<\/p>\n
siege -c 100 -r 20 -b example.com\/path<\/pre>\n
NOTE:<\/strong> We used -d 6<\/code> instead of -b<\/code> in two cases because we wanted to ensure that each server+cache was processing roughly the same number of transactions per second. The -d<\/code> parameter allows you to specify a delay.<\/p>\n
Each test was run five times, and the average result was used for our graphs.<\/p>\n
Web Server Configuration<\/h3>\n
Common settings for LSWS and Apache:<\/p>\n