A Privacy - enhanced Reputation System for Mobile Ad hoc Services

Mobile ad hoc networks (MANETs) are new paradigm of networks oering unrestricted mobility without any underlying infrastructure. Nowadays, MANETs are becoming a promising platform for pervasive social networking and content services. In such services, various information ows over the network. In order to evaluate information trustworthiness, this thesis presents a reputation system to evaluate trust/reputation of entities in MANET services, thus assisting user decisions. We propose AdRep a reputation system with a hybrid trust framework that can support trust/reputation evaluation by both MANET nodes and a trusted server for dierent contexts and scenarios. Thus ephemeral node experiences and/or historical node behaviors can be aggregated to provide more accurate trust information in the context of frequent change of node pseudonyms for privacy enhancement. We further implement two contextual reputation systems AdChatRep and AdContRep, by applying AdRep in two MANET services pervasive social chatting and content recommendation. Our simulation results show AdContRep's eectiveness with regard to unfair rating attacks and on-o attacks. Our two-stage user study evaluates AdChatRep's usefulness, eectiveness and easy acceptance for mobile users. Language: English ii Acknowledgements First and foremost I would like to thank my thesis supervisors, Professor Antti Ylä-Jääski and Professor Øivind Kure. The thesis work would not have been possible without their support, guidance and valuable comments. I owe my deeepest gratitude to my thesis instructor Dr. Zheng Yan for providing me the topic, her shining ideas and patient teaching. Without her help, I could not have learnt so much during thesis work. It is an honor for me to appreciate my colleagues in Nokia who have helped me: It is also a valuable experience to work with them and learn from their brilliant thinking. I am also grateful to all participants in the user study for their precious time, suggestions, comments and encouragement. I sincerely thank my dearest friend Lan Huang who carried out thèGradu-ation Countdown Log' Plan together with me. Her remote spiritual support has helped me remove all obstacles. My sister Yu Su has been constantly caring me during these two years. My atmates Sanja Scepanovic and Heein Song have been accompanying me and bringing so much sunshine to my work and life. My friends as well as classmates Tran Thi Thuy Trang and Jiawei Chen have walked me through diculties during thesis writing. Kiitos! Last but not least, I thank my parents, my brother and my sister-in-law for their eternal love. Thanks …