An Improved Android Collusion Attack Detection Method Based on Program Slicing

Abstract

Android applications can leak sensitive information through collusion, which gives the smartphone users a great security risk. We propose an Android collusion attack detection method based on control flow and data flow analysis. This method gives analysis of data propagation between different applications firstly. And then, a multi-apps program slice model based on both data and control flow are given. Last, the privacy data leakage paths of multi-apps are computed by reaching-definition analysis. Meanwhile, the criterions of mobile device information leakage edge are redefined according to the correlation of mobile devices. Based on the above principle, we implemented an Android collusion attack sensitive information leakage detection tools called CollusionDetector. Case study is carried out for typical collusion attack scenarios and it can obtain better results than existing tools and methods. Experiments show that the analysis of control flow can more accurately find the path of privacy propagation, and more effectively to identify collusion attacks.