[Ns-developers] sta does not receive probe response (from wifi scanning patch)

[ichrak amdouni]

thank you so much for your answer.
I figured out the problem.I had a bug in my code, setting the new channel
twice.
I have noticed a problem similar to what you are describing, it is in the
passive scan, when the switch of the channel and the send of beacons occur
at the same time, and even if the sta receives that beacon later, it
considers the beacon from its new channel although the AP generating the
beacon is on a different channel. I wonder if it is a bug or natural..
That 's said, desynchronizing the two aforementioned events solves the
problem.

Ichrak.







2009/2/17 Gustavo Carneiro <gjcarneiro at gmail.com>

>
>
> 2009/2/3 ichrak amdouni <ichrak.amdouni at gmail.com>
>
> Hi all,
>> I am  using ns3.3 and I have adapted the wifi scanning patch (
>> http://code.nsnam.org/gjc/ns-3-wifi-scanning/rev/d84735b462a8) that
>> Gustavo
>> provided for ns3.2 version.
>> In the scanning procedure, after switching from a channel to another,  the
>> station successfully sends probe requests, APs receive the probe and send
>> probe response, but  the station never receive the probe response from the
>> AP.
>> The issue  is that at the physical layer, station drops packet because
>> 'already in Syn'.
>> Even if I use a unique channel for all nodes (but still calling
>> SetChannel()), I have always the same behavior.
>> If the station does not change the channel, it receives correctly the
>> probe
>> response.
>> Does anyone has an explanation for this.
>> thank you in advance.
>>
>
> Sorry for the delay, but I couldn't think of any answer before.
>
> I just noticed on my own tests that if you have multiple nodes trying to
> scan at the same time some wifi problem (bug or natural, I am not sure)
> occurs due to the synchronization.  Try adding a small random delay to the
> scheduled event that starts the scanning; it will probably solve your
> problem.  It's ok to do this because in the real world independent mobile
> terminals do not usually do things so synchronized.
>
> Hope this helps.
>
> --
> Gustavo J. A. M. Carneiro
> INESC Porto, Telecommunications and Multimedia Unit
> "The universe is always one step beyond logic." -- Frank Herbert
>