A Sensor Network Design from a Probabilistic Automaton Model of Impersonation Attack

We envision a future where thousands to millions of small sensors for self-organizing wireless networks. As it provides solutions to many needs such as wireless communication, remote monitoring, surveillance etc. wireless networking will likely become widespread. Most applications will be done where wiring is impossible to install, or too expensive or where operating and supporting costs are prohibitively high. Thus one need to carefully design a wireless network considering lot of issues like network design, data compression, security parameters etc. We focus on one of such issues of security and provide an estimate of some design parameters of wireless networks. In wireless networks the fundamental nature of communication is broadcast, it is under the control of an adversary. It can modify messages and/or inject fresh messages into the network. It can also try to masquerade as node i and get message from other node. Hence impersonation attack which is a deception whereby one entity purports to be another is frequently occurring one. People have proposed pre-distribution schemes which substantially improve the resilience and security of the network and associated overhead. These schemes rely on probabilistic key sharing among the nodes of a random graph and uses simple protocols for shared-key discovery and path-key establishment. We model an impersonation attack in this framework in terms of a probabilistic automaton. Based on some security parameters we estimate network parameters. Various security and network connectivity parameters are also taken into account in the calculation.