Recent advances in wireless sensor networks (WSNs) are fueling the interest in their application in a wide variety of sensitive settings such as battlefield surveillance, border control, and infrastructure protection. Data confidentiality and authenticity are critical in these settings. However, the wireless connectivity, the absence of physical protection, the close interaction between WSNs and their physical environment, and the unattended deployment of WSNs make them highly vulnerable to node capture as well as a wide range of network-level attacks. Moreover, the constrained energy, memory, and computational capabilities of the employed sensor nodes limit the adoption of security solutions designed for wire-line and wireless networks. In this paper, we focus on the management of encryption keys in large-scale clustered WSNs. We propose a novel distributed key management scheme based on exclusion basis systems (EBS); a combinatorial formulation of the group key management problem. Our scheme is termed SHELL because it is scalable, hierarchical, efficient, location-aware, and light-weight. Unlike most existing key management schemes for WSNs, SHELL supports rekeying and, thus, enhances network security and survivability against node capture. SHELL distributes key management functionality among multiple nodes and minimizes the memory and energy consumption through trading off the number of keys and rekeying messages. In addition, SHELL employs a novel key assignment scheme that reduces the potential of collusion among compromised sensor nodes by factoring the geographic location of nodes in key assignment. Simulation results demonstrate that SHELL significantly boosts the network resilience to attacks while conservatively consuming nodes' resources

Due to the dynamic nature of wireless ad-hoc networks (WAHNs) and the multi-node involvement in most WAHN applications, group key management has been proposed for efficient support of secure communications. Exclusion basis systems (EBS) provide a framework for scalable and efficient group key management where the number of keys per node and the number of re-key messages can be relatively adjusted. EBS-based solutions, however, may suffer from collusion attacks, where a number of nodes may collaborate to reveal all system keys and consequently capture the network. In this paper we investigate the collusion problem in EBS and demonstrate that a careful assignment of keys to nodes reduces collusion. Since an optimal assignment is NP hard, we propose a location-based heuristic where keys are assigned to neighboring nodes depending on the Hamming distance between the strings of bits representing the used subset of the keys.

Given the growing number of applications of wireless sensor networks, some recent research has focused on managing secure wireless communications in such networks. When a large group of sensors that typically are constrained in energy, computation and communication resource are deployed, efficient key management becomes critical. In this paper, we propose a new hierarchical key management scheme for wireless sensor networks based on a combinatorial optimization of the group key management problem. Our solution uses symmetric encryption and re-keying to support current, forward and backward secrecy. An important contribution of our solution is that it yields optimal results for the number of administrative keys per network granule and the number of re-key messages. We also present a mechanism for recovery from a node failure or compromise.