Problem Madwifi - I need 50% packet loss

francisco cardenas <[email protected]> Tue, 15 Nov 2011 17:22:10 +0100
Newsgroups gmane.linux.drivers.madwifi.user
Message-ID <CAC2O_DqDrCHteTnKbaW+nj=Hc4amD+Q+yBXP6H+fXUjOeBPMvg@mail.gmail.com>
--===============5256777484783176940==
Content-Type: multipart/alternative; boundary=f46d043c7fbaa7edab04b1c8636a

--f46d043c7fbaa7edab04b1c8636a
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Dear Sirs;



My name is Francisco C=E1rdenas, I'm a Spanish engineer and I write this em=
ail to
find a solution to my problem with Madwifi.



I am currently working on a research for energy saving in the 802.11
standard and for that I have made a platform with two PCI card (DWL-G520)
with atheros chipset. I also use Madwifi as driver.



The platform consists of the following: a card will act as AP and the other
as station. This station must send a constant number of frames to the
access point (eg, sample temperature). The access point can request a
smaller amount of samples by the beacon frame to the station. To send fewer
frames, I do not under the frecuency of sending of packets in the
estation. What
we really do is lower the transmission power (txpower) to have packet loss
and save energy in the transmission (by IOCTL calls). The problem is that w=
hen
I lose about 20% of packets the card quickly becomes unstable and starts to
have 100% of packets loss. I then dynamically increase power, but takes tim=
e
to stabilize (20 - 30 seconds).



I wish I had a 50% packet loss to increase the margin of my energy saving
algorithm.



My question is if this problem (going from 20% to 100% loss packets) is due
to Madwifi and if a parameter that can be changed.



The funny thing is that when there is a 100% loss and tries to send frames,
the access point is receiving packets. Probably the card is not receiving
acks.



Can you help me?



I would appreciate greatly. Thank you very much for everything.


--=20
Francisco C=E1rdenas

--f46d043c7fbaa7edab04b1c8636a
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable



<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">Dear
Sirs</span></span><span class=3D"shorttext"><span style=3D"mso-ansi-languag=
e:
EN" lang=3D"EN">;</span></span></p>

<p class=3D"MsoNormal"><span class=3D"shorttext"><span style=3D"mso-ansi-la=
nguage:
EN" lang=3D"EN">=A0</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">My
name is</span></span><span style=3D"mso-ansi-language:EN" lang=3D"EN"> <spa=
n class=3D"hps">Francisco</span> <span class=3D"hps">C=E1rdenas</span>, I&#=
39;m <span class=3D"hps">a Spanish engineer</span> <span class=3D"hps">and<=
/span> <span class=3D"hps">I
write this</span> <span class=3D"hps">email</span> <span class=3D"hps">to f=
ind a</span>
<span class=3D"hps">solution to my</span> <span class=3D"hps">problem with<=
/span> <span class=3D"hps">Madwifi</span>.</span></p>

<p class=3D"MsoNormal"><span style=3D"mso-ansi-language:EN" lang=3D"EN">=A0=
</span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">I
am currently</span></span><span style=3D"mso-ansi-language:EN" lang=3D"EN">=
 <span class=3D"hps">working on</span> <span class=3D"hps">a research</span=
> <span class=3D"hps">for
energy saving</span> <span class=3D"hps">in the 802.11 standard</span> <spa=
n class=3D"hps">and for that</span> <span class=3D"hps">I have made</span> =
<span class=3D"hps">a platform with two</span> <span class=3D"hps">PCI</spa=
n> <span class=3D"hps">card</span> (<span class=3D"hps">DWL</span>-G520) <s=
pan class=3D"hps">with</span>
<span class=3D"hps">atheros</span> <span class=3D"hps">chipset</span>. <spa=
n class=3D"hps">I
also use</span> <span class=3D"hps">Madwifi</span> as <span class=3D"hps">d=
river.</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">=A0</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">The
platform</span></span><span style=3D"mso-ansi-language:EN" lang=3D"EN"> <sp=
an class=3D"hps">consists of the following</span>: a card <span class=3D"hp=
s">will act as</span>
<span class=3D"hps">AP</span> <span class=3D"hps">and the other as</span> <=
span class=3D"hps">station. This station</span> <span class=3D"hps">must se=
nd a</span> <span class=3D"hps">constant</span> <span class=3D"hps">number =
of frames</span> <span class=3D"hps">to the access point</span> <span class=
=3D"hps">(eg</span>, sample <span class=3D"hps">temperature).</span> <span =
class=3D"hps">The access point</span> <span class=3D"hps">can request</span=
> <span class=3D"hps">a smaller amount of</span> <span class=3D"hps">sample=
s by</span> <span class=3D"hps">the</span> <span class=3D"hps">beacon</span=
>
<span class=3D"hps">frame</span> <span class=3D"hps">to the station. To sen=
d</span> <span class=3D"hps">fewer frames</span>, I do not <span class=3D"h=
ps">under the</span> <span class=3D"hps">frecuency of sending of packets in=
 the estation</span>. <span class=3D"hps">What we really</span> <span class=
=3D"hps">do is</span> <span class=3D"hps">lower
the</span> <span class=3D"hpsatn">transmission power (</span>txpower) to ha=
ve <span class=3D"hps">packet loss and</span> <span class=3D"hps">save ener=
gy</span> <span class=3D"hps">in the transmission (by IOCTL calls). The pro=
blem</span> <span class=3D"hps">is that</span> <span class=3D"hps">when I l=
ose</span> <span class=3D"hps">about
20%</span> <span class=3D"hps">of packets</span> <span class=3D"hps">the ca=
rd</span> <span class=3D"hps">quickly</span> <span class=3D"hps">becomes un=
stable and</span> <span class=3D"hps">starts to have</span> <span class=3D"=
hps">100%</span> <span class=3D"hps">of
packets loss.</span> <span class=3D"hps">I</span> <span class=3D"hps">then<=
/span> <span class=3D"hps">dynamically</span> <span class=3D"hps">increase<=
/span> <span class=3D"hps">power</span>,
but <span class=3D"hps">takes</span> <span class=3D"hps">time to stabilize<=
/span> <span class=3D"hps">(20 -</span> <span class=3D"hps">30 seconds</spa=
n>).</span></p>

<p class=3D"MsoNormal"><span style=3D"mso-ansi-language:EN" lang=3D"EN">=A0=
</span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">I
wish I had</span></span><span style=3D"mso-ansi-language:EN" lang=3D"EN"> <=
span class=3D"hps">a 50%</span> <span class=3D"hps">packet loss</span> <spa=
n class=3D"hps">to
increase</span> <span class=3D"hps">the margin of my</span> <span class=3D"=
hps">energy
saving</span> <span class=3D"hps">algorithm</span>.</span></p>

<p class=3D"MsoNormal"><span style=3D"mso-ansi-language:EN" lang=3D"EN">=A0=
</span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">My
question</span></span><span style=3D"mso-ansi-language:EN" lang=3D"EN"> <sp=
an class=3D"hps">is if this</span> <span class=3D"hpsatn">problem (</span>g=
oing from <span class=3D"hps">20%</span> <span class=3D"hps">to 100</span>%=
 <span class=3D"hps">loss
packets</span>) is due to <span class=3D"hps">Madwifi</span> <span class=3D=
"hps">and if
a</span> <span class=3D"hps">parameter that can be</span> <span class=3D"hp=
s">changed.</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">=A0</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">The
funny thing</span></span><span style=3D"mso-ansi-language:EN" lang=3D"EN"> =
<span class=3D"hps">is that when</span> <span class=3D"hps">there is a</spa=
n> <span class=3D"hps">100% loss</span> <span class=3D"hps">and tries to</s=
pan> <span class=3D"hps">send
frames</span>, the <span class=3D"hps">access point is</span> <span class=
=3D"hps">receiving
packets.</span> <span class=3D"hps">Probably</span> <span class=3D"hps">the=
 card is not</span>
<span class=3D"hps">receiving</span> <span class=3D"hps">acks</span>.</span=
></p>

<p class=3D"MsoNormal"><span style=3D"mso-ansi-language:EN" lang=3D"EN">=A0=
</span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">Can
you help me?</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">=A0</span></span></p>

<p class=3D"MsoNormal"><span class=3D"hps"><span style=3D"mso-ansi-language=
:EN" lang=3D"EN">I
would appreciate</span></span><span style=3D"mso-ansi-language:EN" lang=3D"=
EN"> <span class=3D"hps">greatly.</span> <span class=3D"hps">Thank you very=
 much</span> <span class=3D"hps">for everything.</span></span><span style=
=3D"mso-ansi-language:
EN-GB" lang=3D"EN-GB"></span></p>

<br clear=3D"all"><br>-- <br>Francisco C=E1rdenas<br><br>

--f46d043c7fbaa7edab04b1c8636a--


--===============5256777484783176940==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
--===============5256777484783176940==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Madwifi-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/madwifi-users

--===============5256777484783176940==--