Distributed Privacy-preserving Techniques For Business-driven Wireless Sensor Network

As the rapid progress in pervasive computing and solid improvement on robustness of wireless data transmission, the mobile data-centric networks, such as Wirelesss Sensor Networks (WSN), Mobile Mesh Networks and Vehicle Ad-Hoc Networks, are stepping into the stage of commercial application development after such a long time for prototype R&D in the previous research history. Among the international academic researches and commercial industries, the issue of network privacy protection has become the primary focus, which is obsolutely the advanced research phases of network security both in terms of theory exploration and application implementation. This thesis makes a through investigation on the privacy protection issues towards Wirelesss Sensor Networks in perspectives of providing the up-to-date literature survey, focusing on the business-driven application, identifying the typical privacy threat in distributed networks, proposing defense techniques as the fundamental mechanism, designing tailored distributed algorithms and conducting experimental simulations. Together with quality of service (QoS) and quality of protection (QoP) in the distributed networking systems, we provide a suite of privacy protection proposal and technuiqe packages. The main contributions enclosed in this thesis are summarized as follows,1) As an emerging research topic in computer society, few papers can provide the latest research progress survey towards the privacy protection problem in WSN. As the first contribution, the thesis categories the up-to-date research progress in terms of the content-aware and context-aware privacy attacks to provide a four-class research framework. In addition, the related works on privacy researches in WSN are also summarized, which have been published in ACM/IEEE top conferences and international journals of computer networking during the latest 4 years.2) As the prevalent knowledge, WSN-based Healthcare Monitoring System application is commercially promising among afore-proposed applications in reality. Based on the practical privacy concerns within the medical data transmission in a Hong Kong local hospital, Multiplex Uniform Attack towards A:-anonymity problem is firstly proposed in this thesis, which considers the essence of the anonymity set in the distributed systems. 3) To take against Multiplex Uniform Attack, k -Anonymity Transfer Mix is thus designed to offer a fundamental mechanism to protect the data privacy of the end users, which can pass the availability testing of four different cost functions. To demonstrate the validity of proposed privacy-preserving mechanism, the mathematical analysis is provided in this thesis, which also shows the seamless compliance with Kerckhoff's principle in security and privacy research academia.4) In addition, 3 distirbitued algorithms are designed to deal with diversified cases in practical networking systems: Fixed k-Anonymity Transfer Mix, Adaptive k -Anonymity Transfer Mix and Lightweight k -Anonymity Transfer Mix. Each of these algorithms has three execution phases, namely logical network initialization, creditable token generation, and privacy-preserving process with different data characteristics and networks.5) Combining the evaluation standards of QoS and QoP, this thesis conducts the simulation experiment in terms of robustness performance, data transmission load and real-time response. For the ease of understanding, the simulation results are presented both in 2D and 3D figures, which contribute more for future work and point out the reliable performance as introduced.