Key Management in Wireless Sensor Networks Using a Modified Blom Scheme

Key establishment between any pair of nodes is an essential requirement for providing secure services in wireless sensor networks. Blom’s scheme is a prominent key management scheme but its shortcomings include large computation overhead and memory cost. We propose a new scheme in this paper that modifies Blom’s scheme in a manner that reduces memory and computation costs. This paper also provides the value for secure parameter t such that the network is resilient. INTRODUCTION Sensor networks comprise a number of sensors with limited resources that are used to collect environmental information and they have been considered for various purposes including security monitoring, target tracking and research activities in hazardous environments. Recent advances in both electronic and computer technologies have increased the demand of wireless sensor networks. Since authentication and confidentiality protocols require an agreed key between the nodes and security of the communication depends on the cryptographic schemes employed, key management is a very important security issue in wireless sensor networks. Earlier key management protocols were based on either symmetric or asymmetric cryptographic functions. Due to resource limitations and lack of security in the sensors, key management protocols based on public keys are inefficient. Hence, symmetric algorithm based key management protocols are used in wireless sensor networks [6]. Several people have proposed different approaches to address the problem of key management [8], [9], [10], [11], [12], [13]. The simple and easy method that could be employed is an online key management center [7]. But, this approach carries a high overhead. Another interesting method is to pre-distribute keys among the sensors, which can result in low cost key establishment in wireless sensor networks. But such schemes fail in handling security or efficiency problems. Some key management protocols use deployment knowledge to increase efficiency. In particular, key agreement schemes for wireless sensor networks must satisfy low energy consumption, low cost, low memory usage, lack of trusted infrastructure, resilient against node capture. Eschenauer and Gligor [1] proposed probabilistic key pre-distribution to establish pair-wise keys between neighboring nodes. In their scheme, each node is preloaded with a key subset from a global key pool in such a way that any two neighboring nodes can share at least one common key with a certain probability. This scheme is vulnerable to the node compromise attack, where keys of normal nodes can be known when some nodes are compromised by adversaries. Chan, Perrig, and Song [2] proposed the q-composite random key predistribution scheme, in which they modified E-G scheme by only increasing the number of keys that two random nodes share from at least 1 to at least q. Their scheme achieves good security under small scale attacks while increased vulnerability in large scale node