[orbit-user] Radiotap header parsing

Wed Aug 26 19:56:34 EDT 2009

Hi Akash,

I made some tests on sb5 at winlab and I made it works using the 
following command line on the node1-1:
trace_oml2 --oml-id d --oml-exp-id e --oml-server tcp:localhost:3003 -i 
int:ath1 -r

Prior to that I've installed the oml-server on the same node but i would 
not recommend it. I had some problems with the ath0 interface so I 
decided to used to ath1 and it work. We need to upgrade the oml-server 
at winlab but I suspect that the previous version is still installed. 
Then, even if the newest version is installed the previous one will 
still be in use. We had this problem once on norbit.

Concerning your previous question about iperf, I have indeed refactorize 
iperf to integrate oml measurement scheme. This new software is called 
iperf_oml2 and is installed with the oml-client package. I will post 
today on http://omf.mytestbed.net/wiki/oml how to use this new software.

Best Regards,

Guillaume

Akash Baid wrote:
> Hi Guillaume,
>
> Thanks for the diagnosis. Clearly, as you said, your table schema and 
> mine is different so we don't have the same version. But I installed 
> every packet related to this application but still I dont get the same 
> schema as yours (One immediate difference is that you have 
> oml_ts_client & oml_ts_server while I have just a single oml_ts). Just 
> to confirm, what I did is:
>
> added deb http://pkg.mytestbed.net/ubuntu hardy/ in the 
> /etc/apt/sources.list
> apt-get update
> apt-get install the following : libpcap0.8, libpcap0.8-dev, libtrace3 
> (libtrace3.0.6) ,libtrace3-dev, liboml, liboml-dev, oml-client. 
>
> I also checked with Ivan to confirm that the console on the sandbox 
> has the latest version of oml-server. I think I am missing some simple 
> point somewhere. Would really appreciate your suggestion as at to what 
> I should try out. 
>
> Thanks
> Akash
>
> On Mon, Aug 24, 2009 at 8:08 PM, Guillaume Jourjon 
> <guillaume.jourjon at nicta.com.au 
> <mailto:guillaume.jourjon at nicta.com.au>> wrote:
>
>     Hi,
>
>     I think you are not using the latest version of both the
>     trace_oml2 and oml2-server during your experiment. Indeed, when
>     I'm running the same experiment using the latest debian packages
>     (oml-client and oml-server) from http://pkg.mytestbed.net/ubuntu I
>     obtain the following results:
>
>     CREATE TABLE trace_oml2_radiotap (oml_sender_id INTEGER, oml_seq
>     INTEGER, oml_ts_client REAL, oml_ts_server REAL, tsft INTEGER,
>     rate INTEGER, freq INTEGER, sig_strength_dBm INTEGER,
>     noise_strength_dBm INTEGER, sig_strength INTEGER, noise_strength
>     INTEGER, attenuation INTEGER, attenuation_dB INTEGER, power
>     INTEGER, antenna INTEGER, sourceMAC TEXT, dstMAC TEXT);
>     INSERT INTO "trace_oml2_radiotap" VALUES(1, 1, 30.6524909995496,
>     0.565649, 1375296379, 2, 2412, -87, -98, 11, 4, 47111, 42447,
>     -125, 1,'00:22:55:ee:b2:90','ff:ff:ff:ff:ff:ff');
>
>     In this example let me explain everything:
>     - oml_sender id = 1 means that this is the first client using oml
>     that connect to the server;
>     - oml_seq = 1 means that this is the first measurement for this table;
>     - oml_ts_client is the timestamp made by the client but
>     re-interpreted by the server that's why you can have some negative
>     values sometimes;
>     - oml_ts_server is the timestamp representing the server time when
>     the server received this measure;
>     - tsft is the Time Synchronization Function timer from the
>     wireless frame in microseconds, it is supposed to be an unsigned
>     64 bits int but we are supporting this type yet. It will be fixed
>     soon;
>     - rate is Tx/RX data rate and the unit is 500Kbps. In this example
>     therefore I have a rate of 1Mbps;
>     - freq is the frequency in MHz, in our example we have a frequency
>     of 2.412 GHz (I'm using 802.11g);
>     - sig_strength_dBm and noise_strength_dBm are the signal and noise
>     strength in dBm;
>     - sig_strength and noise_strength are the signal and noise
>     strength in dB;
>     - attenuation is the wireless transmit attenuation;
>     - attenuation_dB is the wireless transmit attenuation in dB;
>     - power is the transmit power expressed as dBm;
>     - antenna is the unitless indication of the Rx/Tx antenna for this
>     packet.
>
>     I hope that helps you a bit.
>
>     Best Regards,
>
>     Guillaume
>
>
>
>
>     Akash Baid wrote:
>
>         Hi,
>
>         I am using the trace_oml2 application of OML2 for libpcap
>         sniffing. I was wondering if someone can help me out with
>         understanding (parsing) the received radiotap data. The schema
>         and an example row entry from the database is :
>
>         /oml_sender_id INTEGER, oml_seq INTEGER, oml_ts REAL, tsft
>         INTEGER, rate INTEGER, freq INTEGER, sig_strength_dBm INTEGER,
>         noise_strength_dBm INTEGER, sig_strength INTEGER,
>         noise_strength INTEGER, attenuation INTEGER, attenuation_dB
>         INTEGER, power INTEGER, antenna INTEGER, sourceMAC INTEGER,
>         dstMAC INTEGER/
>         /
>         /
>         /"1"// //"16777216"// //"99.76406667032279"//
>         //"-2113757635"// //"369098752"// //"1812529152"//
>         //"-1426063361"// //"-1627389953"// //"201326592"// //"0"//
>         //"0"// //"0"// //"0"// //"33554432"// //"1957724087"//
>         //"1857060791"/
>
>         So I am trying to guess for example how the freq = 1812529152
>         corresponds to channel 1 and so on. Also I am getting some
>         negative numbers for oml_ts, tsft; what is the significance of
>         that? Would really help if someone can point me to the
>         appropriate documentation.
>
>         Thanks
>         Akash
>
>
>
>     _______________________________________________
>     orbit-user mailing list
>     orbit-user at orbit-lab.org <mailto:orbit-user at orbit-lab.org>
>     http://orbit-lab.org/cgi-bin/mailman/listinfo/orbit-user
>
>