Website Fingerprinting by Power Estimation Based Side-Channel Attacks on Android 7

Power consumption data on smartphones can be used to infer user information like web browsing activities and location. Since the power consumption data was traditionally considered harmless to user privacy and developers need to use power information to develop energy-efficient apps, it can be obtained without user permission on Android systems older than Android 7. However, due to the updates in Android 7, the traditional methods of accessing power consumption data either don't work anymore or the data is coarse-grained, and less information can be leaked. In this paper, we propose a power estimation method and design a power estimation based side-channel attack to perform website fingerprinting. Results show that website fingerprinting can be performed with high accuracy using power estimation data on Android 7.