A Security Framework for Cluster-Based Wireless Sensor Networks against the Selfishness Problem

Abstract

Over the last few decades, Cluster-Based Wireless Sensor Networks (CBWSNs) have played a crucial role in handling various challenges (load balancing, routing, network lifetime, etc.) of large scale Wireless Sensor Networks (WSNs). However, the security becomes a big problem for CBWSNs, especially when nodes in the cluster selfishly behave, e.g., not forwarding other nodes' data, to save their limited resources. This may make the cluster obsolete, even destroying the network. Thus, a way to guarantee the secure and consistent clusters is needed for proper working of CBWSNs. We showed that the selfishness attack, i.e., passive attack or insider attack, in CBWSNs can cause severe performance disaster, when particularly a cluster head node becomes selfish. In order to prevent this situation, this paper proposes a security framework that involves a novel clustering technique as well as a reputation system at nodes for controlling selfishness, making them cooperative and honest. The novelty of the clustering comes from the existence of inspector node (IN) to monitor the cluster head (CH) and its special working style. The experimental results showed that the proposed security framework can control the selfishness and improve the security of the clusters.