A Novel Management Model for Dynamic Sensor Networks Using Diffusion Sets

Abstract

The internet of things is predicted to be a complex communication infrastructure embedding millions of devices built around different technologies. It will be using different protocols and operating systems while providing services in different fields. The management of information diffusion in such a complex communication infrastructure is a challenging issue that needs to be addressed efficiently to avoid local disturbances evolve to a large scale disaster by spreading out to the whole infrastructure.This paper revisits the issue of wireless sensor network management to evaluate the performance of information diffusion on connected systems. We propose a novel wireless sensor network partition into sets called "diffusion sets", which depends not only to sensors affinity but also to the mechanisms of the dynamic interactions and the underlying persistent communication model. After proving the partition property, we present the diffusion set computation algorithm and show through experimental results how it can be used by a collection tree algorithm to support efficient network engineering and predictions. Results show that the number of diffusion sets of a network is less correlated with existing network metrics which mater.