Maximum supported network coverage area and cost evaluation of key pre-distribution schemes

  • Authors:
  • Yuan Xue;Helmut Jürgensen

  • Affiliations:
  • University of Western Ontario, London, Ontario, Canada;University of Western Ontario, London, Ontario, Canada and Universität Potsdam, Potsdam, Germany

  • Venue:
  • Proceedings of the 1st international conference on PErvasive Technologies Related to Assistive Environments
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Sensor nodes are low-cost computing devices combining wireless transmitters with specialized sensors, such as temperature, humidity, light, motion, speed, acoustic, chemical and seismic sensors. Sensor nodes are typically deployed in great density over a large geographical area. After deployment, these nodes organize themselves into an ad hoc network to gather the required information from the target environment. A variety of key pre-distribution schemes have been proposed for establishing secure communication in sensor networks. Key pre-distribution schemes bootstrap secret information in each sensor node to enable secure link establishment after deployment. The secure link establishment probabilistically depends on whether two neighboring nodes share some common secret information preloaded before deployment. In this paper, we address key pre-distribution schemes from system aspects: the estimation of maximum supported network coverage area, which memory size to choose and how many nodes to deploy in a target area. Since the unit cost of a sensor node and network density determine the total cost of a sensor network, we believe the analysis of the relationship between memory space, network density and security can benefit the realistic deployment of cost-aware sensor networks. We come to the conclusion that the security performance of key pre-distribution schemes is determined by the cost of the network. The schemes of Eschenhauer et al. and Du et al. can improve the network resilience to node capture or increase the maximum supported network coverage area either by increasing memory space or by increasing network density. On the other hand, in the scheme of Camtepe et al., increasing memory space requires increasing the network density at the same time to improve the security. The scheme of Chan et al. is perfectly resistant to node capture and the maximum supported network coverage area is determined by the memory space only. Contrary to intuition, the scheme of Chan et al. is most efficient when it reduces to preload N -1 keys in each node in the network of N nodes.