Projekte

In cellular networks, the locations of all subscribers are continously tracked even when they only passively carry their mobile devices with them. This privacy sensitive data can be an invaluable source of information, not only for benevolent parties. We therefore present CallForge, the concept of a location management scheme that preserves the subscribers' anonymity -- in many cases even while they participate in a phone call -- as well as a theoretical analysis of the approach.
CallForge improves on PathForge, a previously presented location management scheme, and as such is based on ID switching that we have combined with the emulation of a media break within a single call set-up procedure. We have analyzed and compared the anonymity of PathForge and CallForges, and shown that CallForge consistently provides superior anonymity. Callforge can be implemented entirely in the end device and run on existing network infrastructure without any modifications.

For most mobile networks, the provider needs the current position of their users to provide efficient service. The resulting motion data is not only an invaluable source for analyzing traffic or flow patterns, but also for tracking an individual s whereabouts, even without his knowledge. Today, many carry at least one mobile networked device (cell phone, UMTS-enabled netbook, ...) with them wherever they go, day or night. The resulting motion data can be used to reveal the most intimate details of our lives, making this information extremely privacy sensitive. Nevertheless, little is done to prevent the collection of this information, on the contrary, tracking features and data are being actively collected and even marketed. In this paper, we present PathForge, a lightweight solution, which not only fulfills the provider s efficiency requirement, but continues to allow flow pattern analysis, yet provides full privacy for the user when not actively involved in a call.

Social networks have seen an unprecedented surge of interest in the past few years. Traditionally, they are restricted to central server farms which collect huge amounts of private information from their users. This fails to address two key issues which we expect to shape future social networks: (1) including mobility and location awareness and (2) privacy concerns of the users. Our paper addresses the locality challenge by providing a distributed wireless eer-to-peer infrastructure, which enables discovering of user profiles of nearby users and their friends. More importantly, using easy-to-understand mechanisms that preserve your privacy and render the storage of plain data unnecessary, our system facilitates meeting new friends and recognising old friends in a crowd.
Unlike prior approaches in social networking, we focus on utilising mobile devices that establish direct connections to each other, broadcasting camouflaged information that preserve user privacy without losing the ability of similarity finding using a technology based on a graph representation of a user's data-set and subsequently mapped on a Bloom filter. Furthermore, our approach can be generalised to utilise an inherent property of social networks, namely transitivity, that makes it even more common to get into contact with new, like-minded people.

Social networking is moving to mobile phones. This not only means continuous access, but also allows to link virtual and physical neighbourhood in novel ways. To make such systems useful, personal data such as lists of friends and interests need to be shared with more and frequently unknown people, posing a risk to your privacy. In this paper, we present our approach to social networking, Consistent Deniable Lying (CDL). Using easy-to-understand mechanisms and tuned to this environment, it enables you to meet new friends with joint interests while limiting exposure of your private data. Not only can it be generalised to include friends of friends (transitivity) into interest search, it allows you to plausibly refute any allegations of your claimed interests. Unlike prior work, we focus on the application to similarity finding and include the novel aspects of transitivity and deniability, which are key to success in social networks.