Sinkhole Attack Detection Based On Trust Mechanism In Wireless Sensor Networks

Recent advances in micro-electronic technology, computing and wireless communication have enabled the development of wireless sensor networks (WSNs) that are widely used in military, environmental monitoring areas and so on. However, the security problem of WSNs has received the concern of more and more researchers, which limits the applications of WSNs. There exists many types of attacks threatening WSNs, in which, there is one kind of fundamental attacks-Sinkhole attack. On the basis of this attack, there may be many other attacks deriving from it. By systematically analyzing of existing methods for Sinkhole attacks defense and detection, the paper presents a trust-based multiple point monitoring and acknowledgement mechanism for Sinkhole detection; Combing multiple point monitoring and acknowledgement method with the trust evaluation model, the scheme aims to determine whether there exists sinkhole attack in the networks. We build cooperative a novel trust-based scheme for Sinkhole attacks detection and simulate it in TOSSIM.The highlights and main contributions of the dissertation include:(1) In order to solve the problems of Sinkhole attack prevention and detection mechanism existed in sensor networks, a new attack detection method based on multiple point monitoring and acknowledgement mechanism is put forward. By cooperative detection between sensor nodes and statistical strategy of acknowledge packets, different detection measure is used according to different network status: In routing selection phase, by means of two kinds of packets transmitted between the base station and event area region, statistical node count the number of acknowledge packets to determine whether there exists malicious nodes in the network. While in date transmitting phase, using nodes which received announcement as the monitoring nodes, monitor the process of data forwarding, and send out alarms if the packet loss rate is more than a threshold .Statistical nodes computes the number of alarms to detect Sinkhole attack. The method can achieve relatively high detection rate, consume less resources, avoid complicated encryption and hardware facilities and prolong the life of WSNs. In addition, it is more suitable for practical application.(2) In order to improve the detection rate and avoid the inaccuracy of monitoring of adjacent nodes, in date transmitting phase, we present a trust evaluation scheme which is suitable for WSNs. By systematically analyzing the direct trust degree and the indirect trust degree, we judge if there exits attack in the network. Combined with trust evaluation model, evaluating behavior of nodes, the method can effectively detect Sinkhole attack in the network, and get lower false-negative rate. In addition, it is more suitable for practical application.(3) The process using multiple point monitoring and acknowledgement mechanism for Sinkhole attack detection is simulated in TOSSM using nesC which is a component-based programming language. Then trust-based Sinkhole attack detection mechanism is also simulated. The performance of the secure detection is compared with the existing schemes for Sinkhole attack defense and detection, which validates the validity of the trust-based Sinkhole attack detection mechanism.