[CWN-Summit] Re: Need help gathering evidence

[Nitin H. Vaidya]

.  >>Date: Wed, 2 Feb 2005 12:53:48 -0600
.  >>From: David Young <dyoung at pobox.com>
.  >>
.  >>        * Carefully match sensing range to transmit range, so that
.  >>          "listen before talk" is fair.  It is bad for spectrum
.  >>          sharing if the loudest nodes are also hard of hearing,
.  >>          because then they speak out of turn.  It is also bad if nodes
.  >>          with exquisitely sensitive receivers are always waiting for
.  >>          distant nodes to talk.  Devices that use "listen before talk"
.  >>          should be carefully calibrated at the factory, and/or they
.  >>          should automatically recalibrate their sensing/transmit range.
.  >>          (I am not a big fan of listen-before-talk, but I am a fan of
.  >>          $10 802.11 listen-before-talk radios!)


Here is one way to do this:

Selecting Transmit Powers and Carrier Sense Thresholds for CSMA Protocols,
Jason Fuemmeler, Nitin H. Vaidya and Venugopal V. Veeravalli,
University of Illinois at Urbana-Champaign, Technical Report, October 2004

You can get to this report from
http://www.crhc.uiuc.edu/wireless/groupPubs.html

One can come up with variations on the basic theme in the above report
to trade-off fairness with aggregate throughput.

- nitin