On the privacy of emerging distributed network applications.

摘要

Emerging network applications handle personal data of users as well as sensitive data of intelligent transportation systems. Success of these applications, therefore, relies heavily on their ability to be non-invasive on user privacy while protecting sensitive data. In this dissertation, we investigate user privacy as well as security of some of these applications.; We first develop a much needed framework for secure electronic voting that evaluates the properties of e-voting schemes, and shows the extents to which voter privacy can be guaranteed or sacrificed for an accurate election.; We then investigate privacy issues arising in wireless networks due to unauthorized access to user locations via applications and transmitted signal properties, such as signal strength, enabling location estimation. The resulting location trajectories of a network identifier can be correlated with digital maps to threaten user location privacy (e.g. inferring visited coffee shop or worship place). Specifically, we address location privacy in vehicular ad hoc network (VANET) applications. We demonstrate that predictable mobility of vehicles helps eavesdroppers to track them despite traditional tracking mitigation using identifier updates in the network. As a solution, we propose AMOEBA consisting of the group concept that leverages the group navigation property of vehicles for vehicle-to-infrastructure applications, and the random silent period technique for vehicle-to-vehicle applications.; Privacy, however, is socially and/or culturally defined, and each wireless network user may (not) need privacy at a given location/time. Therefore, in contrast to existing schemes, a user-centric approach is needed with each node determining from available network information, where and when to mitigate its tracking. We design two user-centric schemes: Swing, that enhances privacy gain of the random silent period technique by enabling update when nodes are changing directions; and Swap that can maximize privacy gain using probable exchange of identifiers amongst the updating nodes.; Additionally, in this dissertation we consider protection of sensitive data for next-generation network-enabled airplane models, such as 787, that will use network applications for decreasing costs and easing maintenance. We focus on the security of a specific application, i.e. the electronic distribution of safety-critical loadable software and data, and propose a standardized security framework.