jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andreas Fransson <Andreas.Frans...@se.flextronics.com>
Subject What sort of deviation is it?
Date Fri, 12 Sep 2003 12:37:39 GMT
OK listen up,

I've been struggeling with JMeter for a while and done some tests, but one
thing still confuses me. You get nice looking graphs containing Data,
Average and Deviation. I suppose that Data stands for a sample's Round Trip
Time (RTT) and Average for the average RTT value, BUT Deviation? What kind
of deviation is it. I have hard to believe that is it standard deviation due
to some factors. But the most obvious one, if you spam a server until is
starts dropping packets the Average and Deviation grows bigger, right? (In
most casese anyway) If the Deviation is Standard Deviation how on earth can
the same amount of samples appear below the average value as above. Example:
(the number are just made up for visual purpose)

Data:	Average:	Diviation:
200	106		106
98	106		106		
100	106		106
99	106		106
99	106		106
300	106		106
98	106		106
98	106		106

This is a scenario that happens most of the time. The Data values are
concentrated round the average and some samples takes way longer time than
the average one. Please say that the diviation isn't a standard diviation. I
would like to use a chi-square distribution to approx. the samples. If you
know what distribution that is used, please let me know...

//Andreas


Mime
View raw message