Modeling Cascading Failures for Wireless Sensor Networks With Node and Link Capacity

Abstract

Existing cascading models for wireless sensor networks (WSNs) mainly focus on the impact of node capacity on network performances and fail to consider the link capacity. Moreover, the traffic metrics they used cannot properly reflect the sink-convergence feature of WSNs. In this paper, we build a more practical cascading model for WSNs, in which the network is limited by node capacity and link capacity, and the network load is defined according to a new traffic metric “directional edge betweenness.” In addition, failed nodes are allowed to recover after a certain time delay, instead of being permanently deleted from the network. To improve the network invulnerability against cascading failures, a routing recovery mechanism is presented. Simulation results show that the network invulnerability is closely related to cache resources and bandwidth resources owned by sensor nodes. There exist thresholds for link capacity and node capacity that can make the network reach the steady state.