[laa-wifi-simple] Cannot observe duty cycle pattern and its impact on Wi-Fi

[Xuhang Ying]

Dear all,

I am playing around with laa-wifi-simple and want to study LTE-DC and Wi-Fi coexistence, but cannot interpret the results I got. 

I used the following command:

$ ./waf --run "scratch/laa-wifi-simple --lteDutyCycle=0.5 --pcapEnabled=true --duration=3 ChannelAccessManager=DutyCycle"

I suppose that LTE eNB will be transmitting for 50% of the duty cycle (e.g., 80ms by default I suppose), and turned off for the rest of time. So Wi-Fi nodes should be able to communicating during the OFF time, not during the ON time. I expected to see the recurring ON/OFF pattern by looking at Wi-Fi AP/Station traffic (laa-wifi-ap-1-1.pcap, laa-wifi-sta-3-1.pcap). But I saw something really different (see attached screenshot), and I don't quite understand it.

Any ideas?

Thanks,

Shaun

[Tommaso Pecorella]

Hi,

the fact that the Wi-Fi nodes should send packets in the off periods doesn't mens that they must not try to do it during the on periods.

The duty cycle in LTE will only means that the LTE nodes will blindly transmit packets following a given pattern. if the Wi-Fi nodes will try to send stuff when the channel is used by LTE they'll just collide.

If you have a collision or a channel busy detection... well, that's another problem.

T.

[Xuhang Ying]

Hi Tommaso,

Thanks for your post. Suppose that Wi-Fi nodes are very close to LTE-U eNB. Since Wi-Fi nodes employ CSMA/CA, they are supposed to sense ongoing LTE transmissions via energy detection in CCA, and thus defer their transmissions during ON periods. Isn't it not the case? That is what I was hoping to see from the LAA WiFi example.

Best,

Shaun

[Tommaso Pecorella]

Hi,

well, it depends also on the LTE activity. Transmissions during the active periods are still subject to the fact that there's something to transmit :)

T.