An Efficient Algorithm for Cut Vertex Detection in Wireless Sensor Networks

Due to lack of precise control on network topology, cut vertices often exist in wireless sensor networks (WSNs). A cut vertex is defined as a sensor node whose removal breaks network connectivity. Failures of such nodes make a network disconnected and block data transmission, hence it is vital to detect cut vertices in WSNs. This paper proposes a distributed algorithm CVD to identify cut vertices. The algorithm travels the nodes of a network in parallel, colors the edges based on the interval-coded spanning tree, and then determines the cut vertices by counting the edge colors. The parallel style of the algorithm significantly reduces the time delay. Further more, optimization strategies based on interval coding are proposed to reduce the communication cost of the algorithm. The correctness of the algorithm is proved and its performance is analyzed theoretically. Theoretical analysis shows that it requires O (dn) communication cost and has O (dn) time delay, where n is the number of nodes in a WSN, and d is the maximum degree of sensor node. Both the simulation results and the experiments on real sensor networks show that the algorithm outperforms previous algorithms with lower communication cost and shorter time delay.