Andro Lyze: A Distributed Framework for Efficient Android App Analysis

Abstract

In recent years, the number of mobile applications has grown significantly. Not surprisingly, various security and privacy concerns associated with mobile applications have emerged. Several researchers are addressing these problems by analyzing the security properties of mobile application code. Most of the security checks rely on custom scripts and are quite heterogeneous with respect to dependencies, deployment and reporting. In this paper, we present AndroLyze, a distributed framework with unified logging and reporting functionality to perform security checks on large numbers of applications in an efficient manner. AndroLyze provides optimized scheduling algorithms for distributing static code analysis tasks across several machines. Moreover, AndroLyze can handle several versions of a single mobile application to generate a security track record over many versions. To demonstrate the benefits of AndroLyze, we have analyzed the Top Free 500 Android applications of all categories in Google Play collected over three years. The resulting data set consists of almost 40,000 mobile applications and requires about 227 GB of storage space.