Preferential Detachment: Improving Connectivity and Cost Trade-offs in Signaling Networks
NECSI and Brandeis University
NECSI and University of Massachusetts Dartmouth
Sandia National Laboratory
Last modified: April 24, 2006
We introduce a distributed, non-hierarchical adaptive method for significantly improving network transmission and connectivity in a signaling network of autonomous nodes with distance-dependent links. Using the framework of multihop wireless networks, we achieve preferential detachment of links in regions of high node density through an adaptive algorithm that selectively decreases an individual node’s broadcast radius. Due to fewer incoming links, more simultaneous signals are tolerated by the network, thus offsetting the increase in path length. With cost savings as high as 90% for various spatial topologies requiring full internodal communication, our method highlights the role of self-adjusting node transmission power not only in minimizing simultaneous signaling, but also in minimizing overall power consumption.