We are looking at using Neoload for testing the effectiveness of a number of SSL offloading / acceleration solutions.
However after setting up Neoload to perform some basic initial testing, the results show no improvement at all in CPU utilisation of the target system when SSL acceleration is enabled vs running the exact same tests when acceleration is disabled.
Basic OpenSSL benchmarks performed on the target VM show the acceleration is actually working OK.
We have read through the answers forums and it appears that Neoload proxy (where the SSL client connections are spawned) are 'optimised' which we are thinking may be the reason why Neoloads tests are not be giving us 'true real world' SSL client behaviour.
We understand that perhaps most people using Neoload are doing so in order to test their application and NOT ssl offload performance, and as such would like the SSL component 'filtered out' - however we are actually wanting to measure the effectiveness of SSL offloading systems (and not actually so much the application) and as such we need SSL client connections to be setup and handshakes as 'realistically' as possible.
Could you please let us know why our Neoload tests appear to be disabling SSL acceleration and what are the recommended settings are in order to set Neoload up specifically to test SSL handshake and de-crypt performance?
Kind regards,
Jeff