Research On Group Key Management Technology In Wireless Sensor Networks

Wireless sensor networks (WSNs) have been widely used in both military and civilian field such as battlefield surveillance, habitat monitoring, healthcare and traffic control. Due to insecure wireless channels and constrained resources, WSNs are susceptible to various kinds of attacks. Security becomes extremely important especially when WSNs are deployed in hostile environments. As a result, secure communications in some wireless sensor networks are critical. Among all the basic research field of secure wireless sensor networks, key management is the fundamental security mechanism. Group key is one of the most important key management paradigms for group communication which is both bandwidth-efficient and energy-efficient. Hence, establishing a common group key for the whole network is a vital concern to prevent an adversary from attacking group communication.This paper systematically investigates some key issues of key management in wireless sensor networks, especially on the key distribution security and performance, which are the most important requirements of key management schemes. The research mainly focuses on the following three aspects.First, we propose a novel lightweight key management based on EBS and polynomial in wireless sensor networks. A dynamic key management scheme based on the exclusion basis system(EBS) was proposed by researchers to minimize the number of keys and reduce the number of re-keying messages in group communications. EBS is a combinatorial optimization methodology which approves to be secure and efficient for key management of group communication setups. The drawback of EBS is that it could not resist collusion attack. In the worst case only two collusive nodes are required to compromise all keys in the network. To address collusion problem, this paper develops a novel method to combine EBS with broadcasting polynomials. EBS works to refresh keys when there is no collusion attack, otherwise broadcasting polynomials are used to re-key the network. Our scheme is secure when there are no more than t compromised nodes in the network. Performance analysis shows that our scheme is communication and storage efficient with desirable scalability.Second, we design and analyze a routing based two-level key management scheme in wireless sensor networks. On the basis of the existing tree-based key pre-distribution schemes, this paper proposes a new hierarchical key management scheme in combination with the routing topology in WSN. Considering different communication requirements among sensor nodes, the inter-cluster protocol takes re-encryption and authentication mechanisms while the intra-cluster assigns an administrative node for generating and maintaining the routing key tree. The results show that this new scheme can highly achieve secure and efficient data transmission. Third, we address the key management in WSNs with unreliable wireless channel and propose a Limited Self-Healing Key Management(LiSH) with collusion-resistant and revocation ability. Self-healing key distribution is a good choice for group communication in WSNs with unstable and lossy wireless channel. Many self-healing schemes have been proposed in recent years. In this paper, we propose a new self-healing key distribution with t-revocation capability. Considering security requirement and practical utility, we limit the number of lost keys that could be retrieved from broadcast message. In addition, we develop a secure self-healing mechanism using one-way function to protect current session keys from being attacked by intruders. A dirty window attack is also defined in our scheme and we provide a simple yet efficient method to avoid it. Security analysis shows that our scheme is collusion-free and achieves t-wise forward and backward security. Comparison with other schemes indicates that our scheme is more efficient in terms of storage and communication complexity.To summarize, this paper investigates two issues of key management: security and performance, which are key requirements in wireless sensor networks. The theoretical analysis and simulation verification all show that our proposed schemes can properly achieve the design goals.