|Abstract:||Wireless sensor networks (WSNs) were originally motivated by military applications, and are becoming integral part of more and more civilian applications to improve quality of life. With current wireless sensor network technology, people will gain advanced knowledge of physical and social systems, and the advent of a ubiquitous sensing era is coming. In-network processing or data aggregation is an essential function of WSNs to collect raw sensory data and get aggregated statistics about the measured environment, and help queriers capture the major feature or changes of the measured systems. As more and more applications of WSNs collect sensitive measurements of people's everyday life, privacy and security concerns draw more and more attention.
If privacy of sensory content is not preserved, it is not feasible to deploy the WSNs for information collection. On the other hand, if integrity of the collected sensory information is not protected, no queriers or users can trust and/or use the collected information. Hence, two important issues should be addressed before wireless sensor network systems can realize their promise in civilian applications: (1) protect data privacy, so the deployment of the wireless sensor network systems is feasible; (2) enforce integrity, so users can trust the collected information (or aggregated result). This dissertation explores privacy and integrity of data aggregation in wireless sensor networks.
First, I present two privacy-preserving data aggregation schemes for additive aggregation functions, and show that the additive aggregation functions can serve to estimate the aggregation results for more general aggregation functions. The first scheme, Cluster-based Private Data Aggregation (CPDA), leverages clustering protocol and algebraic properties of polynomials. It has the advantage to enable peer monitoring within a cluster. The second scheme, Slice-Mix-AggRegaTe (SMART), builds on slicing techniques and the associative property of addition. It has the advanii tage of incurring less computation overhead for privacy-preserving data aggregation.
Then, I address both privacy of individual sensory data and integrity of aggregation result simultaneously. It is very challenging to achieve the synergy of privacy and integrity, because privacy-preserving schemes try to hide or interfere with data, while integrity protection usually needs to enable peer monitoring or public access of the data. Therefore, privacy and integrity can be the conflicting requirements, one may barricade the implementation of the other. I extend SMART and CPDA to preserve privacy and make the queriers able to verify the integrity of data aggregation.
To show the efficacy and efficiency of the proposed schemes, I present simulation results of our schemes and compare their performance to a typical data aggregation scheme, Tiny Aggregation protocol (TAG), where no privacy preservation and integrity protection is provided. We explore multiple dimensions in design space, and investigate the trade-offs in protocol design. To the best of our knowledge, this dissertation is among the first network protocols to preserve privacy and integrity in data aggregation for wireless sensor networks.