IEEE 802.11be (Wi-Fi 7)simulation and caluculation Throughput
jomon charly
NS_FATAL, terminating
terminate called without an active exception
Command 'build/scratch/ns3.42-wifi-eht-networkd-default' died with <Signals.SIGABRT: 6>.
Tom Henderson
Dear Sir/Madam,
I am Experiencing two problem with the simulation of Wi-Fi 7 in ns 3.42.
Problem 1When i was try to associate more than 50 station with a AP(without MaxMissedBeacons attribute) I am getting an error like this.
NS_ASSERT failed, cond="GetMac()->CanForwardPacketsTo(rxAddr)", msg="Cannot forward frame to 00:00:00:00:00:1f", +6.684031000s 62 file=/home/jomon/ns-allinone-3.42/ns-3.42/src/wifi/model/wifi-mac-queue-scheduler-impl.h, line=369
NS_FATAL, terminating
terminate called without an active exception
Command 'build/scratch/ns3.42-wifi-eht-networkd-default' died with <Signals.SIGABRT: 6>.
If MaxMissedBeacons attribute is set to 8, most of the time i am getting Throughput as Zero. Sometimes I am getting the same error.
What will be the optimal value of MaxMissedBeacons.
When you have 50 to 100 stations on a single AP, you need to take care that they all successfully associate. Can you check that they are all successfully associating? The error that you cite can be caused by not associating successfully, or by losing the association.
Rather than guessing, you can explicitly check on associations
and de-associations to see if that is the problem. There are
trace sources for Assoc and DeAssoc in the StaWifiMac.
When you start an ns-3 simulation with a lot of clients, all STAs are not associated at time zero. Then, the AP sends a beacon, and all STAs try to associate at the same time, which leads to a lot of collision failures. Then, if some STAs do successfully associate and start sending traffic, they can occupy the channel to the extent that the other STAs who failed to associate early are blocked.
Therefore, it is a good idea to explicitly check that all STAs are associated before starting to send data. This is what I call a warmup period. If you have 50 STAs, for example, hook the Assoc and DeAssoc traces and wait until you have 50 associations and no de-associations. Then, at that simulation time, you are ready to start sending data.
For an example program that uses these traces to log all
associations and de-associations, please see wifi-bianchi.cc in
the directory src/wifi/examples/
Problem 2
As per theory throughput must be inversely proportional to the number of stations. But in some values of number of stations, throughput values are higher compared to lower number of station(i have done the experiment multiple times to validate)
(Our setup consists of one access point, 5-100 stations, and four links/channels (one 2.4 GHz, one 5 GHz, and two 6 GHz links). We have configured constant mobility and placed stations 100 meters apart, with a simulation time of 100 seconds.)
Please anyone help me in this.
If you have separation of nodes that varies up to 100 meters, some of the longer range connections will not be able to use a high MCS (although the shorter range connections will be able to do so). That could be a source of the results not matching with your expectations.
I would recommend that you try first to place all of your nodes close together (say, within 10 meters) and see if the larger distances are the problem. Then, if not, I think that you should reduce the number of nodes and increase from a small amount to see where the results diverge from your expectation.
- Tom