The impact of non-repudiation and preserving perfect sink unobservability on the lifetime of wireless sensor networks

In energy-limited wireless sensor networks, first three of confidentiality, integrity, authentication and non-repudiation security services are typically implemented by either using pure symmetric crypto primitives or with a hybrid public key and secret key algorithm combination. Non-repudiation requirement has a unique characteristic in the sense that it can be met only by digital signatures which is usually implemented with public key algorithms. Hence this service has been traditionally considered to be infeasible in wireless sensor networks. In the first part of this thesis, we construct a linear programming framework to analyze the impact of implementing a non-repudiation service on the lifetime of wireless sensor networks. Feeding the model with published results of RSA and ECDSA algorithms and parameters derived for a modified communication-efficient version of one-time signatures, we compare these algorithms with respect to their effect on network lifetime in different topologies and operating conditions. We have shown that with proper algorithm selection and for a wide range of network parameters, the lifetime reduction due to the utilization of digital signatures can be less than 10 % and 20 % for security levels of 80-bit and 112-bit respectively compared to a network not utilizing digital signatures.In wireless sensor networks an attack to the base station (sink) can render the whole network useless hence concealing the physical location of the sink may be necessary in certain circumstances. Previous studies addressing this challenging problem do not study in the presence of an eavesdropper who has global knowledge for the entire network. In the second part of this thesis, a naive solution is to employ fake sinks so that nodes send their data not only to the real sink but also other locations. An alternative solution with less overhead could be provided when all nodes including base station equalize the values of their total incoming and outgoing flows as well as their energy expenditure. By this way, no information about sink location is revealed when all communication within the network is monitored. . In the second part of this thesis, through a Linear Programming framework we analyze and compare impact of these solutions on the lifetime of wireless sensor networks.