A key management based on multiple regression in hierarchical sensor network

To lead a present communication paradigm to ubiquitous world, sensor networking is a core technology. Especially guaranteeing secure communication between sensor nodes is critical in hostile environments, and key management is one of the most fundamental security services to achieve it. However, because the structure of sensor networks can be very various according to their application, there can not exist the best solution for all applications. Thus, we design a key management scheme on hierarchical sensor network, to take advantage of the topology. To support both scalability and resilience against node capture, we apply a multiple regression model to key generation, calculation and extension. The proposed scheme is based on the key pre-distribution, but provides the key re-distribution method for key freshness. To overcome the weakness of centralized management, the role of key management is partially distributed to aggregators as well as a sink. These management nodes need not store keys except them for re-distribution, and can calculate them easily using key information from nodes, as needed. Performance results show that the proposed scheme can be applied efficiently in hierarchical sensor network compared with other key managements.