Controlling inter-arrival times

[Phil Steitz <ph...@steitz.com>]

Is it possible in Jmeter to develop tests that are controlled not by the 
number of "concurrent threads" simulating "concurrent users" but rather 
by a pre-determined distribution of inter-arrival times?

What I mean is the following.  In our high-volume web applications, we 
see a widely fluctuating number of "concurrent connections" as back end 
service times vary and users spend different amounts of time on the 
site.  Tests based on ramping up and maintaining "concurrent 
connections" do not give an accurate picture of how our applications 
actually perform under load. To simulate production load in a load test, 
we would like to be able to let the thread count go up and down as 
necessary to accomodate a distribution of inter-arrivals that is either 
a) random, exponentially distributed with mean = some fraction of 
measured peak mean inter-arrival time or b) based on actual replay or 
resampling from production inter-arrival times.

Is it possible to do any of this using jmeter?

If not, is there a straightforward way to modify jmeter to support this 
capability.

thanks,

Phil


---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org