Fine-grained Dynamic Privacy Leakage Analysis Of Android Third-party Libraries

With the development of mobile Internet,mobile terminal devices are widely used in our daily life.The large-scale use of Android has promoted the number of Android applications increase exponentially,meanwhile,the security issues of applications emerge in an endless stream.It is difficult to carry out an unified and effective behavior monitoring on the complex Android applications in the application market.Among them,a massive amount of personal privacy data is stored in the Android mobile terminal,including contacts,SMS and so on.With the development of the big data technology,Android terminal system is becoming the main channel to obtain the personal privacy information.Today,mobile users are paying more attention to mobile application security and personal privacy.But,more worryingly,for the question "the users' privacy information is actually acquired by whom,and how it is used",users can never get a clear answer.In view of the common privacy leakage problems in Android system and applications,this paper refines the research granularity of privacy leakage analysis.According to the different developers,an Android application is considered to be divided into the host application and multiple third-party libraries.The privacy leakage risks of third-party libraries are uncontrollable to both of the application developers and users,which should be the focus of the privacy leakage flows.On the basis of theoretical research,this paper puts forward two different Android privacy leakage analysis tools,providing a fine-grained dynamic analysis technology.The main contents of this thesis are as follows:(1)Based on network flow and reverse analysis,this paper studies the Android application's source code,analyzes the privacy leakage characteristics of Android apps,and identifies three types of privacy leakage path of the host application and third-party libraries inside apps,which is the theoretical basis for privacy leakage research.(2)Combing of static third-party library detection tool and dynamic Xposed framework,this paper proposes a fine-grained and dynamic privacy leakage analysis tool to analyze the privacy leakage behaviors of third-party libraries in real time.(3)To track the privacy flows more comprehensively,according to the privacy leakage features of Android system,this paper studies deeply from the system perspective,a fine-grained bi-directional tracking privacy leakage analysis system is designed and implemented,which is called "FineDroid".FineDroid tool defines the propagation process of privacy flows,tracks and analyzes the whole process of privacy information receiving,transmitting,leaking and the return flows' receiving and spreading.The analysis process distinguishes different subjects,such as the host application and the third-party libraries.Based on the actual propagation process information of privacy flows,by matching with the fine-grained privacy leakage path models,the privacy leakage risk of each privacy flow is dynamic evaluated in multi-angle and full-scale.As for the two analysis tools designed in this paper,the lightweight plug-in based on Xposed is easy to use,no need to modify the Android system and applications,and the resource consumption is small;the analysis based on the Android system bottom frame of the system level FineDroid tool is more truthful and reliable.The multi-dimensional evaluation system for the privacy leakage is more consistent with the actual application scenario,and it has a high practical value.The two tools can meet the needs of a variety of application scenarios.The feasibility and effectiveness of our tools are verified by experiments of large-scale application detection.Our study effectively solves the problem of " users' privacy information is acquired by whom and how it is used".