Research On Replication Attack Detection Of Mobile Wireless Sensor Network

Wireless sensor networks are often arranged in extremely hostile environments, where the adversary is easily able to capture one or more sensor nodes. Once the sensor nodes are captured, an attacker can re compile the nodes by using the reverse technique and replicate a large number of nodes with the same identity information. These nodes are arranged in the network because of their legal status, it is easy to be considered as a legitimate node of the security protocol. They are active in the network through a variety of means of destruction, and ultimately achieve the purpose of subversion or control network. The detection of node replication attacks in a wireless sensor network is therefore a fundamental problem.Currently, replication attack detection scheme for wireless sensor network is divided into two major categories:centralized and distributed. The advantages of centralized detection scheme is low storage overhead and high detection rate. As a result, the centralized detection scheme can not be separated from the base station, so the problem of single point failure is the biggest problem. Secondly, the traffic flow of the base station is relatively large compared with other regions, which leads to the unbalanced load in the network and reduces the life cycle of the network. Centralized detection scheme can be understand easily and has good detection success rate. Distributed detection scheme does not need such a node, so the current research direction is mainly focused on the distributed detection scheme. The existing distributed detection scheme is divided into two categories, the deterministic detection scheme and the random detection scheme. The determination method has the advantage of low storage cost, but the detection effect of the protocol is easy to be destroyed by the capture of verification node from the attacker. The random detection scheme, although the security of the detection scheme is higher, but also with a higher computational overhead. Among all of the security challenges of wireless sensor networks, replication attack can be a highly damaging attacks which can subvert the network. This kind of attack use the sensor node which has been captured to create a large number of nodes that have the same identity information of the captured one, to theft of the information, tamper with the information, phagocytosis data and even to achieve the purpose of network paralysis. We propose a protocol by detecting the count of the unique secret evidence of the node and the number of the shared key in the network to achieve the purpose of detecting replication nodes. After the theoretical analysis and simulation experiments we provide our protocol can has good feasibility and effectiveness in the mobile wireless sensor network replication attack. At the same time, this paper gives a theoretical guidance to the number of the number of the shared key and the count of the unique secret evidence.