Opportunistic computing in wireless networks

Opportunistic computing is the exploitation of opportunities as they arise to provide computing services that meet application requirements. It maximally utilizes available resources to improve such services subject to application constraints. When applied to different applications, we have different concrete instances of opportunistic computing. In this dissertation, we focus on its utilization in the context of wireless networking applications. Particularly, we study four instances of opportunistic computing in wireless networks: opportunistic social networking, opportunistic localization, opportunistic encryption, and opportunistic authentication.;First, we study opportunistic social networking with mobile phones. We develop E-SmallTalker, a novel distributed mobile communication system that aims to facilitate more effective social networking among strangers in physical proximity. We propose a privacy-preserving opportunistic commonality discovery protocol that utilizes iterative Bloom filters to encode user information. We eliminate unnecessary user interactions by exploiting Bluetooth service attributes to publish encoded user data.;Second, we study opportunistic localization in wireless sensor networks. We propose an "anti-sensor network" system to localize an adversary's sensors in a noncooperative environment where these sensors try to evade or deter localization. The system relies on a set of monitors opportunistically observing intermittent wireless signals emitted by sensors in the protected area. With these observations, we estimate sensor locations with an opportunistic localization algorithm. We also design improved algorithms to handle additional countermeasures that sensors can employ such as message encryption and non-uniform transmission power levels.;Third, we study opportunistic encryption in wireless LANs. We identify and analyze security issues in the current 802.11i security standard. We illustrate the severe consequences resulting from the lack of frame authentication with an exemplary denial-of-service attack. We propose a new key-establishment algorithm for opportunistic encryption in the link layer. Our algorithm makes minimal changes to the existing 802.11 protocol.;Fourth, we study opportunistic authentication in wireless web applications. We propose opportunistically utilizing users' trusted devices such as mobile phones to aid web authentication in wireless networks. We design and implement a prototype one-time-password (OTP) authentication system that works seamlessly in heterogeneous environments. We propose a one-time-password reference service that allows an OTP-token to be opportunistically used in multiple web applications. We also propose a new connection-aware one-time-password algorithm to thwart man-in-the-middle attacks by using connection information.