[copy to the White Rabbit Developers mailing list] I am considering distributed radiofrequency signal acquisition and generation using two Ettus Research X310 SDR platforms, fed with the 1 PPS and 10 MHz outputs from our White Rabbit (WR) network. In addition to these signals, the control software (libuhd) running on the computer collecting data from all SDR platforms needs to timestamp packets in order to synchronize acquisitions. If I run a 1GbE network through a switch, all works fine, I can collect synchronous data from the local and remote site. In real life I'd like to use the WR network to transfer the data: I know my network configuration is good since I can ping the USRPs. I have increased MTUs on the host Ethernet interfaces and the WRS (running firmware 5.0.1) to 9000 as advised by libuhd. In all cases the synchronization of the USRPs fail with error messages UHD Error: x300 fw communication failure #1 EnvironmentError: IOError: x300 fw poke32 - reply timed out while the same configuration works when replacing the WR network with a 1GbE switch. Is there some sort of excess latency introduced by the networking layer of WR that might be the cause of the timeout ? Is there any configuration I can tune to try to solve the issue ? Or lower some of the timeout limit on libuhd ? Thanks, Jean-Michel

Hi Jean-Michel, As you hinted, it might be best to re-post this to the White Rabbit list: https://lists.ohwr.org/sympa/info/white-rabbit-dev >From my limited experience with White Rabbit and the N310 product, I don't see any obvious issues, other than possibly within the switch. -Daniel On Wed, Oct 17, 2018 at 1:17 PM jean-michel friedt via USRP-users < usrp-users@lists.ettus.com> wrote: > [copy to the White Rabbit Developers mailing list] > > I am considering distributed radiofrequency signal acquisition and > generation using > two Ettus Research X310 SDR platforms, fed with the 1 PPS and 10 MHz > outputs from our > White Rabbit (WR) network. In addition to these signals, the control > software (libuhd) running > on the computer collecting data from all SDR platforms needs to timestamp > packets in order > to synchronize acquisitions. If I run a 1GbE network through a switch, all > works fine, I can > collect synchronous data from the local and remote site. In real life I'd > like to use the > WR network to transfer the data: I know my network configuration is good > since I can ping > the USRPs. I have increased MTUs on the host Ethernet interfaces and the > WRS (running > firmware 5.0.1) to 9000 as advised by libuhd. > In all cases the synchronization of the USRPs fail with error messages > UHD Error: > x300 fw communication failure #1 > EnvironmentError: IOError: x300 fw poke32 - reply timed out > while the same configuration works when replacing the WR network with a > 1GbE switch. Is there > some sort of excess latency introduced by the networking layer of WR that > might be the cause > of the timeout ? Is there any configuration I can tune to try to solve the > issue ? Or lower some > of the timeout limit on libuhd ? > > Thanks, Jean-Michel > > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com > -- Daniel Jepson Digital Hardware Engineer National Instruments O: +1.512.683.6163 daniel.jepson@ni.com