SpotME: promoting location privacy, one lie at a time
Last week at ICDS and today at Eurecom, I presented our work on location privacy. Here is the basic idea -
By sharing their location on mobile social-networking services, mobile phone users benefit from a variety of new services working on *aggregate* location data such as receiving road traffic estimations and finding the best nightlife "hotspots" in a city. However, location sharing has caused outcries over privacy issues - you cannot really trust private companies with your private location data ;) That's why we have recently proposed a a piece of software for privacy-conscious individuals and called it SpotME (here is the paper). This software can run directly on a mobile phone and reports, in addition to actual locations, a very large number of erroneous (fake) locations. Fake locations:  are carefully chosen by a so-called randomised algorithm, guarantee that individuals cannot be localized with high probability, yet they have little effect on services offered to car drivers in Zurich and to subway passengers in London. For technical details, please have a go at the paper ;)
ICDCS 2011
just before the workshop in france, i attended icdcs in the states. few papers follow:
Efficient and Private Access to Outsourced Data (pdf). Say that you outsource your private data to "the cloud". The authors of this paper proposed a new data strucutre with which you can efficiently access your outsourced data while guaranteeing content, access, and pattern confidentiality from any observer, including the cloud provider.
Dissecting Video Server Selection Strategies in the YouTube CDN (pdf). Ruben presented an extensive study of the YouTube CDN. The goal of this study was to identify the factors that impact how video requests are served by data centers. They found that "the YouTube infrastructure has been completely redesigned compared to the one previously analyzed in the literature. In the new design, most YouTube requests are directed to a preferred data center and the RTT between users and data centers plays a role in the video server selection process. More surprisingly, however, our analysis also indicates a significant number of instances (at least 10% in all our datasets) where videos are served from non-preferred data centers."
An Energy-efficient Markov Chain-based Randomized Duty Cycling Scheme for Wireless Sensor Networks. Giacomo presented a new duty cycling scheme for sensor nodes that is energy-efficient and is based on Markov chains. In the past, Giacomo has done some interesting work in the area of "Internet of Things" at Sun Labs: he built a Web-based application that analyses and visualise large, heterogeneous, and live data streams from a variety of devices (pdf).
Efficient Online WiFi Delivery of Layered-Coding Media using Inter-layer Network Coding (pdf). Dimitrios studied the problem of how to deal with the problem of client diversity when video is multicasted to multiple clients over a wireless LAN. He showed that the traditional triangular scheme for inter-layer network coding performs poorly. He thus proposed a new online video delivery scheme that can be deployed behind the wireless AP.