Privacy-preserving framework for context-aware mobile applications

In recent years, the pervasiveness of mobile devices, especially mobile phones and personal digital assistants (PDAs), has increased rapidly. At the same time, wireless communication networks have improved considerably and the usage of mobile devices to access the internet is, with decreasing costs, possible almost everywhere and at any time in industrialized countries. However, the usage of mobile technology and mobile applications to support business processes, trans- actions, and personal tasks is still low compared to their potential. The improved capabilities resulted in the introduction of many applications for mobile devices by network operators and software vendors. These services were meant to increase the average revenue per user (ARPU) on top of the voice call income. But many of these services have failed and none of them has led to an improved usage of mobile services today, besides e-mail. A new kind of application, the context-aware application, exploits the ubiquity of the mobile devices in order to fit the personal need or task the user is about to execute satisfactorily. Context-aware systems try to improve the communication with the user by adding information about the current context to the explicit user input and by adapting the output to the current setting of the user. While those applications are seen as important steps to a widespread usage, there are strong factors inhibiting their development and adoption. First of all, the lack of common frameworks handling context data and improv- ing software development increases the cost to build context-aware applications. Each application currently implements its own sensors and logic to handle its data. Furthermore, service providers need to offer tailored services for every con- text of the user. Since no single provider is able to be an expert for all kinds of applications and will not have the necessary number of developers, a common service which finds services of multiple providers for the current situation of the user is needed. All services need to utilize the context attributes which are locally determined by the user’s situation. Development costs are further boosted by the difficulty of developing applications for multiple devices with varying input/output (IO) capabilities like speech output, small and big screens, full qwerty-keyboards, touchscreens, or numeric keypads. From the user’s perspective, privacy also endangers the adoption of mobile services. Context information may include very private data and expose the user’s preferences and habits. While the user may trust a single, well-known, provider to secure the private data and to respect the user’s privacy concerns, the problem increases with more and more smaller service providers