Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

Twinkle: Network Power Scheduling in Sensor Networks

Barbara Hohlt and Eric A. Brewer

EECS Department
University of California, Berkeley
Technical Report No. UCB/CSD-05-1409
April 2005

http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/CSD-05-1409.pdf

We present our experience with Twinkle, the first implementation of network-layer power scheduling with real applications. Twinkle uses dynamically created schedules to schedule network flows in sensor networks. The scheduling allows nodes to turn off their radio when idle, thus saving power. Twinkle supports broadcast and partial flows for flexibility, and integrates time synchronization to enable scheduling to work on real motes. We show that it avoids contention, increases fairness, and that it can achieve power savings of 2-5x for real applications over existing power-management schemes, and over 150x compared with no power management. Finally, we discuss the extensions to the original algorithm needed for real applications.


BibTeX citation:

@techreport{Hohlt:CSD-05-1409,
    Author = {Hohlt, Barbara and Brewer, Eric A.},
    Title = {Twinkle: Network Power Scheduling in Sensor Networks},
    Institution = {EECS Department, University of California, Berkeley},
    Year = {2005},
    Month = {Apr},
    URL = {http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/6472.html},
    Number = {UCB/CSD-05-1409},
    Abstract = {We present our experience with Twinkle, the first implementation of network-layer power scheduling with real applications. Twinkle uses dynamically created schedules to schedule network flows in sensor networks. The scheduling allows nodes to turn off their radio when idle, thus saving power. Twinkle supports broadcast and partial flows for flexibility, and integrates time synchronization to enable scheduling to work on real motes. We show that it avoids contention, increases fairness, and that it can achieve power savings of 2-5x for real applications over existing power-management schemes, and over 150x compared with no power management. Finally, we discuss the extensions to the original algorithm needed for real applications.}
}

EndNote citation:

%0 Report
%A Hohlt, Barbara
%A Brewer, Eric A.
%T Twinkle: Network Power Scheduling in Sensor Networks
%I EECS Department, University of California, Berkeley
%D 2005
%@ UCB/CSD-05-1409
%U http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/6472.html
%F Hohlt:CSD-05-1409