Re: Measure the latency for a specific throughput/datarate method

["Tiago Junqueira" <tiago.junkeira --at-- gmail.com>]

Thanks a lot Jeff,

 In fact when i was analysing the server status using iperf i saw:
tiago --at-- ksat67:~$ iperf -c 192.168.101.11 -u -b 500k -t 30
[  3]  0.0-10.0 sec    479 KBytes    393 Kbits/sec  9.184 ms    0/  334 (0%)
[  3] 10.0-20.0 sec    471 KBytes    386 Kbits/sec  11.203 ms   83/  411 (20%)
[  3] 20.0-30.0 sec    468 KBytes    383 Kbits/sec  10.314 ms   99/  425 (23%)
[  3]  0.0-32.5 sec  1.50 MBytes    388 Kbits/sec  11.123 ms  205/ 1277 (16%)

So the final value 11.123 ms is not the average of the other ones. So
iperf just return the Jitter(Last) as the overall jitter that is
already a "moving average", without making any statical treatment,
giving it just as it is. But as you said this is not a good way to
measure the overall jitter, is more important for real time
applications, to see the real time jitter and maybe compensate.

     Thanks for the explanation, it's more clear for me now.