Dear, Deeply thank you for your reply. My working environment is based on iperf3 version. There is a one server and one client in my testing environment. Both of them are running with Ubuntu Mate OS version (32 bit). The server is Intel NUC core i7 and the client is Rapsberry Pi 3 Model B. The testing network type is software defined wireless network. RYU controller is installed at the server side. The aim of using TCP iperf3 is testing the TCP wireless link throughtput between Server(where RYU controller is installed) and client. There is no middle box in my case as the client and server are directly connected. For my case, I have observed that TCP Iperf3 is stopped working because the network interface is congested. I have to restart both server and client if I want to do another iperf3 testing. The window size is default in the experiment.
Best Regards, Soe Ye Htet On Tue, Sep 18, 2018, 4:59 AM Bruce A. Mah <b...@es.net> wrote: > If memory serves me right, Soe Ye Htet wrote: > > > I have a problem with using TCP IPerf testing on my simple wireless > > network. There are one server and one client. The command of " iperf -s > > -i 10" is applied at the server side and the command of "iperf -c > > <server ip address> -t 10000 -i 10" is applied at the client side. The > > reason for putting a high number of seconds is aimed for the reliability > > testing. However, the connection between server and client was broken > > down after 1 hour which is 3600 seconds. I tried to debug for the > > connection problem but I could find the fact of faulty in my experiment. > > If anyone knows, please suggest me which part do I need to troubleshoot > > for this issue. > > Is this for iperf version 2 or iperf version 3? (The main reason I am > asking is that it looks like you commented on someone else's issue in > the iperf3 issue tracker.) > > iperf3 does not have any intrinsic limit on the total duration of a > test, and as far as I know, iperf2 doesn't either. I'm wondering if > maybe you have some middlebox (or host-based firewall) that is imposing > a maximum duration on the TCP connection being used between the two > hosts? Is that possible? > > Bruce. > >
_______________________________________________ Iperf-users mailing list Iperf-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/iperf-users
