[Dataset] rutgers/capture (v. 2007-04-20)

top

the initial version

@MISC{rutgers-capture-2007-04-20,
  author = {Kishore Ramachandran and Marco Gruteser and Ivan Seskar and Sachin Ganu and Jing Deng},
  title = {{CRAWDAD} data set rutgers/capture (v. 2007-04-20)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/rutgers/capture},
  month = apr,  
  year = 2007
}
					

In an experiment involving two senders and one receiver, 
we placed a sniffer (wireless NIC in monitor mode) close to each of 
the senders so as to capture all transmitted MAC frames from each sender.

We experimentally investigate the physical layer capture
effect in off-the-shelf 802.11 network cards and confirm that
it reduces throughput fairness of traffic flows. 

All our experiments were conducted on the ORBIT testbed comprising
64 wireless nodes arranged in an 8x8 grid. Each node has two 802.11 
a/b/g cards. We used 802.11b channel 1 for all our experiments. There 
is an equal distribution of nodes with Intel IPW 2915 chipset based cards 
and Atheros AR5212 chipset based cards.

For all our experiments, we have used the nodes with Atheros
cards since they allow software control over various parameters
such as CWmin selection, disabling retries etc. The open source
Madwifi driver for the Atheros chipset based cards implements
a majority of MAC protocol features in the driver rather than in
hardware, thereby allowing a variety of modifications at the software
level. We have also developed a supporting software library that 
allows us to extract useful information such as RSSI, PHY rate,
hardware timestamp (1μsecond granularity) from the device driver
for each successfully received packet. Note that there are no hidden
nodes in our testbed and each node is within transmission range
of every other node. There is no external interference from other
802.11 wireless devices in all our experiments. This was verified
by using the iwlist (interface) scan utility that detects infrastructure
or ad-hoc networks in the vicinity.

To experimentally detect the physical layer capture phenomenon,
we adapted the technique of using per-sender sniffers and constructing
a global timeline of all transmission and reception events in each of 
our experiments.

[Traceset] rutgers/capture/RFMON (v. 2007-04-20)

top

the initial version

@MISC{rutgers-capture-RFMON-2007-04-20,
  author = {Kishore Ramachandran and Marco Gruteser and Ivan Seskar and Sachin Ganu and Jing Deng},
  title = {{CRAWDAD} trace set rutgers/capture/RFMON (v. 2007-04-20)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/rutgers/capture/RFMON},
  month = apr,  
  year = 2007
}
					

We used a wireless NIC in RFMON mode (sniffer) close
to each of the senders so as to capture all transmitted
MAC frames from a particular sender.

In these experiments, we use two transmitters S1 and S2 that
send packets to a common receiver. We chose one sniffer near each
sender such that the signal strength or RSSI of packets received from 
this sender is higher than that of frames received from any other sender. 
The reasoning behind this placement is that a sniffer is also a regular 
radio receiver susceptible to the capture phenomenon. We use a feature 
provided by Atheros cards - a station can perform "live monitoring" and 
observe WLAN traffic while still being synchronized with the rest of
the stations in the network. This implies that the logs from each of
the sniffers do not have to be explicitly "synchronized"; they can
be merged directly based on the hardware timestamp of each received
frame. We used tcpdump on the sniffers and processed the collected 
information using awk scripts.

[Trace] rutgers/capture/RFMON/sender1 (v. 2007-04-20)

top

the initial version

@MISC{rutgers-capture-RFMON-sender1-2007-04-20,
  author = {Kishore Ramachandran and Marco Gruteser and Ivan Seskar and Sachin Ganu and Jing Deng},
  title = {{CRAWDAD} trace rutgers/capture/RFMON/sender1 (v. 2007-04-20)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/rutgers/capture/RFMON/sender1},
  month = apr,  
  year = 2007
}
					

We used a wireless NIC in RFMON mode (sniffer) close
to each of the senders so as to capture all transmitted
MAC frames from a particular sender.

Traces included:
----------------

We used a wireless NIC in RFMON mode (sniffer) close
to each of the senders so as to capture all transmitted
MAC frames from a particular sender. This trace contains
tcpdump file from sniffer close to sender 1.

Further processing:
-----------

Step 1: You can first convert the traces to text using
tethereal (tshark) as follows:

% tethereal -Vr sender1.trace > sender1.txt

Step 2: You can then extract relevant information from the
each trace using the included awk script:

% awk -f process-tcpdump.awk sender1.txt > sender1-processed.txt
% awk -f process-tcpdump.awk sender2.txt > sender2-processed.txt

Step 3: You can then merged the traces using the UNIX sort utility. 
Note that since the sniffers were also part of the 802.11b infrastructure 
network, (we were leveraging a monitor mode provided by madwifi which 
allowed us to do so) the wireless NIC at the sniffers is synchronized 
using access point beacons. Thus, we can merge the trace by sorting 
on the 64-bit receive timestamp provided by Atheros hardware (with 
microsecond granularity). More details regarding the setup are available 
in our workshop paper.

% sort -n -k 1 sender1-processed.txt sender2-processed.txt > merged-trace.txt

This trace (tarball) includes an awk script (process-tcpdump.awk) and a tcpdump format (including Prism Monitoring Header)

[Trace] rutgers/capture/RFMON/sender2 (v. 2007-04-20)

top

the initial version

@MISC{rutgers-capture-RFMON-sender2-2007-04-20,
  author = {Kishore Ramachandran and Marco Gruteser and Ivan Seskar and Sachin Ganu and Jing Deng},
  title = {{CRAWDAD} trace rutgers/capture/RFMON/sender2 (v. 2007-04-20)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/rutgers/capture/RFMON/sender2},
  month = apr,  
  year = 2007
}
					

We used a wireless NIC in RFMON mode (sniffer) close
to each of the senders so as to capture all transmitted
MAC frames from a particular sender.

Traces included:
----------------

We used a wireless NIC in RFMON mode (sniffer) close
to each of the senders so as to capture all transmitted
MAC frames from a particular sender. This trace contains
tcpdump file from sniffer close to sender 1.

Further processing:
-----------

Step 1: You can first convert the traces to text using
tethereal (tshark) as follows:

% tethereal -Vr sender1.trace > sender1.txt

Step 2: You can then extract relevant information from the
each trace using the included awk script:

% awk -f process-tcpdump.awk sender1.txt > sender1-processed.txt
% awk -f process-tcpdump.awk sender2.txt > sender2-processed.txt

Step 3: You can then merged the traces using the UNIX sort utility. 
Note that since the sniffers were also part of the 802.11b infrastructure 
network, (we were leveraging a monitor mode provided by madwifi which 
allowed us to do so) the wireless NIC at the sniffers is synchronized 
using access point beacons. Thus, we can merge the trace by sorting 
on the 64-bit receive timestamp provided by Atheros hardware (with 
microsecond granularity). More details regarding the setup are available 
in our workshop paper.

% sort -n -k 1 sender1-processed.txt sender2-processed.txt > merged-trace.txt

This trace (tarball) includes an awk script (process-tcpdump.awk) and a tcpdump format (including Prism Monitoring Header)

[Author] Kishore Ramachandran

top

[Author] Marco Gruteser

top

[Author] Ivan Seskar

top

[Author] Sachin Ganu

top

[Author] Jing Deng

top

[Paper] ganu-capture

top