I am struggling to solve an issue on our point to point links where WLAN interfaces connect with 500-650Mbps but throughput test with /tool traffic-generator indicated capacity less than 100Mbps average although all ethernet interfaces are transmitting and receiving at 1Gbps.
See below of tx/rx rates on station-bridge side on station-bridge device which is immediately in front of the traffic generator source.
Below is also the results of the traffic-generator tool quick test
Monitoring of traffic on wlan device close to the traffic generator source indicated limitation 100Mbps from ether1 to wlan1 interfaces although ether1 is a gigabit interface. Since wlan1 is receiving at less than 100Mbps then there is no way the other side of the link wlan1 can receive above 100Mbps.
My question is there a limitation in bridge switch chip transmission between ether1 and wlan1.
Are you aware that ROS built-in traffic generator is heavy on CPU and might load CPU to 100% … both decreasing actual wireless performance (which is always somehow CPU-bound) and showing unrealistically low result (because it’s traffic generator which can’t generate higher traffic, not interface not being able to pass it). So always monitor CPU usage when running traffic generator …
Hi
The traffic generator source device is RB3011 and the gateway is also CCR device. The wireless ptp link are DISC Lite5 ac devices less than 3km apart.
Both the traffic generator source device and intermediate wlan ptp devices have low cpu usage (less than 5%) during test.
