A Dynamic Clustering Wireless Sensor Network Key Management Scheme

The security of wireless sensor networks(WSN)is directly related to the further application of IoT,and it is also one of the important technical means for IoT to acquire data information.The sensor nodes typically have characteristics of limited battery energy,weak storage capacity,limited computing ability and low communication bandwidth,these features give more difficulty to the design of key management scheme in WSN which requires not only less complexity and fewer resources consumption,but also be able to ensure the safety and reliability of the network.Therefore,traditional key management scheme cannot be applied to WSN directly.This paper proposes a dynamic key management scheme for hierarchical wireless sensor networks based on the complexity of solving elliptic curve discrete logarithm problem.In particular,the main contributions of this thesis are list as follows.1.Aiming at the vulnerability of keys and the complexity of key management due to dynamic node changes in wireless sensor networks,this scheme employs asymmetric key system for node authentication between cluster head node and base station node in the key agreement process.The cluster key management is based on the Shamir's(t,n)threshold secret sharing mechanism between common nodes and cluster heads,this scheme divides the key into n sub-keys and distributed to each node,and less than t compromise nodes won't release the message of the master key.2.Aiming at the existing random key pre-distribution management scheme is limited by the node storage space and other issues,this scheme generates a key based on the public key combination system on the elliptic curve,and generates a large number of keys through a combination of key matrix elements with less information.Each sensor node only needs to store the relevant parameters and the key matrix,thereby greatly saving the node memory space.3.Aiming at the problem of unbalanced network energy consumption due to unreasonable number and distribution of cluster heads in cluster-based key management schemes,this paper proposes a multi-criterion approach for the selection of cluster heads(CHs)using Technique for Order Preference by Similarity to Ideal Solution(TOPSIS)to reduce network energy consumption caused by cluster head selection.Four attributes are considered for the selection of CHs,namely residual energy,number of neighbors,distance from the base station and the transmission range for each node.Based on these four criteria,a set of cluster heads were selected by Pareto optimal solution,and then ordered by TOPSIS method.Analysis and comparison result show that the scheme has strong safety,it can resist fake node attacks,message replay attacks,denial of service attacks and other common attacks in WSN,and have a greater advantage in network connectivity rate,storage efficiency,network energy consumption and other aspects,compared with the traditional of the programs.It can be applied to large hierarchical wireless sensor networks.