[Tinyos-help] TOSSIM Problem

Philip Levis pal at cs.stanford.edu
Sun Sep 4 16:53:27 PDT 2005


On Sun, 2005-09-04 at 20:08 -0230, Brett Parsons wrote:
> Hi Philip,
> 
> Yes, you are correct!  I misread the TOSSIM documentation.  It should 
> have been obvious considering they weren't in transmission range of each 
> other!  I'm going to add a small, random delay before broadcasting 
> messages.  Would you recommend another approach or should this suffice?

There's actually a good deal of research on this topic. Approaches
include:

1) Rebroadcast with some probability p, which is some function of the
density. How you decide what p is depends on your cost/benefit trade-
offs.

2) Wait a random time t in some interval [a,b], rebroadcast then. The
range of |b-a| should also be dependent on the density (what works well
for a sparse network may overload a very dense one).

3) Same as 2), except reset your timer if you hear a broadcast.

4) If you want reliable delivery, then a single broadcast (or, really,
any finite number of transmissions) per node is not sufficient, as
there's some probability that a node won't hear any of them. There are
ways to control your broadcast rate to achieve rapid propagation and low
transmission rates when everyone has the newest message.

Some good papers to look at for these approaches:

1) Bimodal Multicast, Kenneth Birman et al.
2) The Broadcast Storm Problem, Sze-Yao Ni et al.
3) I don't know of any specific ones, but it was used in Design and
Implementation of a Sensor Network System for Vehicle Tracking and
Autonomous Reception, Cory Sharp et al.
4) A Self-Regulating Algorithm for Code Propagation and Maintenance in
Wireless Sensor Networks, Philip Levis et al.

Phil



More information about the Tinyos-help mailing list