Hardware-Assisted Two-Party Secure Computation on Mobile Devices

This thesis focuses on the practical realization of general two-party Secure Function Evaluation in a mobile environment and the possibility of enhancing these techniques by the use of a trusted hardware token. Secure function evaluation allows multiple mutually distrusting parties to jointly compute a function on their private inputs without revealing anything but the function output. This technique is particularly interesting in the context of mobile electronics, such as smartphones, where typically highly sensitive user data is stored and processed. The protection of this data is desirable but very costly, due to the high complexity of secure computation protocols. Implementing Secure Function Evaluation schemes on smartphones is a challenging task due to their limitations in processing power, memory and battery-life. To address these issues, we extended an existing two-party secure function evaluation scheme by a trusted hardware token that allows to securely pre-generate data, used in the actual function evaluation phase for masking sensitive values. For the purpose of securely distributing data generated by the token, we designed and implemented a communication protocol based on TLS on the smart card. We present working demonstrators for managing the hardware token and running secure two-party function evaluation on Android smart phones making use of a microSD smart card. The use cases we implemented are private set intersection to find shared contacts and securely scheduling a meeting. Our implementation is benchmarked and its performance is analyzed.