Wireless Sensor Network Key Management Scheme

Wireless sensor networks (WSN) have an open architecture, the physical security in the distribution area can not be guaranteed, therefore, it is particularly important to achieving effective secure infrastructure, and key management scheme is crucial technologies. Since the computation capacity and energy source are limited in WSN, most conventional schemes can not be used in WSN directly; it makes a challenge to achieving effective secure infrastructure in WSN. Researchers at home and abroad do much work on key management scheme in WSN, but there are still some issues, and deserve further study.This dissertation focuses on improving security, reducing the storage overhead and communication overhead in WSN, and proposes two kinds of key management schemes.(1) The proposed key management schemes for WSN are mostly based on symmetric cryptography, with poor performance on resistance to node capture. To solve this problem, a key management scheme based on Elliptic Curve Cryptography is proposed. Hash function is adopted to generate the pairwise key. Use Elliptic Curve Cryptography to provide security of key parameters, improve the security of the keys by one-way operation of Hash function. In order to validate such positive effects of the proposed scheme, simulations are performed on the connectivity, security performance, storage overhead and energy consumption. The results of simulations demonstrate that the proposed scheme has better performances to improve the resilience to node capture and brute-force attack.(2) There is a contradiction between connectivity and security in typical key pre-distribution schemes. To solve this problem, a new key pre-distribution scheme based on linear combination of vector group is proposed. The proposed scheme pre-distributes a vector to each node, and generates the pairwise keys using this vector. The pairwise keys and cluster keys could be generate by one-time broadcasting, so the proposed scheme can effectively support the sub-cluster network. Simulations on security performance, storage overhead and energy consumption are performed to validate the positive effects of the proposed scheme. The results of simulations demonstrate that the proposed scheme could reduce the storage overhead and energy consumption, and improve the resilience to node capture.The first proposed scheme has improved the security by using asymmetric cryptosystem, which can meet the higher security requirement. The other is the improvement of typical random key-predistribution schemes, it has improved the security, reduced the storage overhead and energy consumption, which is suitable for the restricted resource environment.