Skip to content

Performance

sneakybeaky edited this page Apr 9, 2012 · 3 revisions

The logic in mod_redis is mainly concerned with applying the config rules to work out what redis operation to perform and the encoding of the response, so most of the performance is down to the latency to the redis server and the redis operation itself.

Here's some figures generated by the JMeter test in the repo run on a 2.53Ghz Intel Core 2 Duo MacBook Pro running OS X 10.6.8 with 4GB memory, Apache 2.2.22 and Redis 2.4.10. This test simulates 50 concurrent users executing 1,000 scenarios with a ramp up period (initial pause between each user) of 1 second. The scenario each user uses is randomly chosen - either a simple create, read and update of a simple value or a form post to update a sorted set.

Aggregate Report

Transactions per second :

Transactions per second

Response times :

Response times

Clone this wiki locally