Skip to first unread message
saumil shah
Mar 2, 2018, 10:31:14 AM3/2/18
to ns-3-users
Dear All,
I am working on ns-3-lbt module. I have some doubts as mentioned below.
Case 1 : CellconfigA=Lte CellconfigB=Wifi and lteDutyCycle=0.5
Case 2: CellconfigA=Laa CellconfigB=Wifi and --ChannelAccessManager=DutyCycle with Onduration=40ms and Onperiod=80ms
As mentioned above both cases try to implement 50%duty cycle for LTE. But there is a big difference between delay at IP layer in both case.
Case 1 : ./waf --run "laa-wifi-simple --cellConfigA=Lte --lteDutyCycle=0.5 --d2=100 --ChannelAccessManager=DutyCycle --duration=5"
--------monitorA----------
Flow 1 (1.0.0.2:49153 -> 7.0.0.2:9) proto UDP
Tx Packets: 31437
Tx Bytes: 47155500
TxOffered: 75.4488 Mbps
Rx Bytes: 25153500
Throughput: 40.2456 Mbps
Mean delay: 395.452 ms
Mean jitter: 0.562976 ms
Duration: 5.35992 s
FirstTX pkt Time: 2.06408 s
LastRx pkt Time: 7.424 s
Rx Packets: 16769
Lost Packets: 14668
Case 2 : ./waf --run "laa-wifi-simple --cellConfigA=Laa --d2=100 --ChannelAccessManager=DutyCycle --duration=5"
with Onduration=40ms and Onperiod=80ms
--------monitorA----------
Flow 1 (1.0.0.2:49153 -> 7.0.0.2:9) proto UDP
Tx Packets: 31437
Tx Bytes: 47155500
TxOffered: 75.4488 Mbps
Rx Bytes: 27817500
Throughput: 44.508 Mbps
Mean delay: 1330.13 ms
Mean jitter: 0.451775 ms
Duration: 5.89592 s
FirstTX pkt Time: 2.06408 s
LastRx pkt Time: 7.96 s
Rx Packets: 18545
Lost Packets: 12892
As you can see in case 1 delay is 395 ms and in case 2 delay is 1330 ms which is quite high.
In both cases we are trying to use 50% duty cycle then why there is such a high delay in case 2?
Best Regards
Saumil Shah
I am working on ns-3-lbt module. I have some doubts as mentioned below.
Case 1 : CellconfigA=Lte CellconfigB=Wifi and lteDutyCycle=0.5
Case 2: CellconfigA=Laa CellconfigB=Wifi and --ChannelAccessManager=DutyCycle with Onduration=40ms and Onperiod=80ms
As mentioned above both cases try to implement 50%duty cycle for LTE. But there is a big difference between delay at IP layer in both case.
Case 1 : ./waf --run "laa-wifi-simple --cellConfigA=Lte --lteDutyCycle=0.5 --d2=100 --ChannelAccessManager=DutyCycle --duration=5"
--------monitorA----------
Flow 1 (1.0.0.2:49153 -> 7.0.0.2:9) proto UDP
Tx Packets: 31437
Tx Bytes: 47155500
TxOffered: 75.4488 Mbps
Rx Bytes: 25153500
Throughput: 40.2456 Mbps
Mean delay: 395.452 ms
Mean jitter: 0.562976 ms
Duration: 5.35992 s
FirstTX pkt Time: 2.06408 s
LastRx pkt Time: 7.424 s
Rx Packets: 16769
Lost Packets: 14668
Case 2 : ./waf --run "laa-wifi-simple --cellConfigA=Laa --d2=100 --ChannelAccessManager=DutyCycle --duration=5"
with Onduration=40ms and Onperiod=80ms
--------monitorA----------
Flow 1 (1.0.0.2:49153 -> 7.0.0.2:9) proto UDP
Tx Packets: 31437
Tx Bytes: 47155500
TxOffered: 75.4488 Mbps
Rx Bytes: 27817500
Throughput: 44.508 Mbps
Mean delay: 1330.13 ms
Mean jitter: 0.451775 ms
Duration: 5.89592 s
FirstTX pkt Time: 2.06408 s
LastRx pkt Time: 7.96 s
Rx Packets: 18545
Lost Packets: 12892
As you can see in case 1 delay is 395 ms and in case 2 delay is 1330 ms which is quite high.
In both cases we are trying to use 50% duty cycle then why there is such a high delay in case 2?
Best Regards
Saumil Shah
saumil shah
Mar 5, 2018, 9:30:07 AM3/5/18
to ns-3-users
Dear All,
In addition to above post i have similar question about delay of LAA node using LBT mechanism.
I have produce the results based on laa-wifi-simple.cc script. In my opinion i think delay of LAA node is quite high and in real world application it can not be acceptable. I have run the scripts with different configuration as mentioned below.
Case 1: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Laa --d2=10 --duration=5"
In addition to above post i have similar question about delay of LAA node using LBT mechanism.
I have produce the results based on laa-wifi-simple.cc script. In my opinion i think delay of LAA node is quite high and in real world application it can not be acceptable. I have run the scripts with different configuration as mentioned below.
Case 1: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Laa --d2=10 --duration=5"
------
Throughput: 41.3352 Mbps
Mean delay: 1423.24 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 42.0984 Mbps
Mean delay: 1411.96 ms
------
Case 2: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Laa --d2=10 --duration=10"
Throughput: 41.3352 Mbps
Mean delay: 1423.24 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 42.0984 Mbps
Mean delay: 1411.96 ms
------
Case 2: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Laa --d2=10 --duration=10"
------
Throughput: 37.3356 Mbps
Mean delay: 2527.47 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 39.138 Mbps
Mean delay: 2436.6 ms
------
Case 3: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Laa --d2=10 --duration=20"
Throughput: 37.3356 Mbps
Mean delay: 2527.47 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 39.138 Mbps
Mean delay: 2436.6 ms
------
Case 3: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Laa --d2=10 --duration=20"
------
Throughput: 36.0198 Mbps
Mean delay: 4709.86 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 36.7224 Mbps
Mean delay: 4483.05 ms
As we can see from above 3 cases that in all 3 cases both cells are configured as LAA and just duration is changing. As we increase the duration mean delay is also increasing.
First question : Are theses delay normal? I think it is very high delay.
Second question : Why with increasing the simulation duration delays are also increasing ? It might change but that should be very small and not this much difference.
Case 4: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Wifi --d2=10 --duration=20"
Throughput: 36.0198 Mbps
Mean delay: 4709.86 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 36.7224 Mbps
Mean delay: 4483.05 ms
As we can see from above 3 cases that in all 3 cases both cells are configured as LAA and just duration is changing. As we increase the duration mean delay is also increasing.
First question : Are theses delay normal? I think it is very high delay.
Second question : Why with increasing the simulation duration delays are also increasing ? It might change but that should be very small and not this much difference.
Case 4: ./waf --run "laa-wifi-simple --cellConfigA=Laa --cellConfigB=Wifi --d2=10 --duration=20"
------
Throughput: 55.8612 Mbps
Mean delay: 2373.18 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 12.4116 Mbps
Mean delay: 381.208 ms
In Case 4 i have done simulation with one wifi cell and one LAA cell. In this case also we are getting delay of LAA cell is 2373.18 ms which is high. I think in real world application 2.3 seconds of delay will not be acceptable.
Can someone please comment about above why are we getting high delay ? Is this normal or am i doing something wrong or is my understanding is wrong?
Best Regards
Saumil Shah
Throughput: 55.8612 Mbps
Mean delay: 2373.18 ms
------
--------monitorB----------
Flow 1 (2.0.0.2:49153 -> 7.0.0.3:9) proto UDP
------
Throughput: 12.4116 Mbps
Mean delay: 381.208 ms
In Case 4 i have done simulation with one wifi cell and one LAA cell. In this case also we are getting delay of LAA cell is 2373.18 ms which is high. I think in real world application 2.3 seconds of delay will not be acceptable.
Can someone please comment about above why are we getting high delay ? Is this normal or am i doing something wrong or is my understanding is wrong?
Best Regards
Saumil Shah
Reply all
Reply to author
Forward
0 new messages