Research On Investigation Of Privacy Protection In Android Operating System Based On Static Analysis

There is a rapid growth in smartphone market in the past few years. While smartphones become more and more popular and powerful, they take over many works that used to be done by other devices, such as taking photo, navigation, paying bill, surfing the internet, sending email, etc. During use, users would save a lot of private data in their phone, including photos, emails, documents, browsing history, etc. It is an important task for the smartphone operating system to protect these private data efficiently. As the most popular operating system of smartphone, Android enforces a permission-based security policy to protect the private data. Nowadays, there is significant research interest in the security policy of Android system.In this paper, we propose an investigation method based on static analysis, to investigate the effectiveness of privacy protection in Android system. After identifying all possible interfaces that can be exploited by malicious apps to read or send the private data, we investigate each interface to check if it can read or send private data, behind users’ back. The main works in this dissertation are listed as follows:1. Propose a formal method based on static analysis, to investigate the effectiveness of privacy protection in Android system. From an attacker’s perspective, the formal method can detect vulnerabilities of the privacy protection mechanism in Android.2. Propose an attack graph generator to extract attack graph from the privacy protection mechanism in Android. After identifying the private data and the sensitive interfaces reading or sending the private data, the attack graph generator extracts the attack graph by analyzing the execution paths for every sensitive interface. 3. Implement a logic analyzer to analyze the attack graph of private protection mechanism, which is described as Horn clauses. By traversing all possible attack paths in the attack graph, the logic analyzer can find all possible valid attack paths, which can be exploited by malicious apps to read the private data and then send them out, without requesting any related permission.4. Reveal several vulnerabilities in the privacy protection mechanism of Android. After being applied to12revisions of Android system, from Android2.1to the recently released4.4.2, the investigation method reveals several vulnerabilities in the privacy protection mechanism.8categories out of22of private data can be accessed by malicious apps without any related permission. The vulnerabilities are verified in8Android devices, the results indicate that these vulnerabilities have existed for years; almost all Android devices are impacted.5. Propose several possible solutions to fix the vulnerabilities in Android system, and several methods to detect the malicious apps exploiting these vulnerabilities.6. Implement a malicious app and publish it to Google Play. The experimental results indicate that users will get no sensitive warning from Google Play when they are installing this app, and get no security alert from several popular security tools when this app is reading or sending the private data.7. Propose an optimization method to alleviate the problem of high false positive rate in static analysis during the investigation. After analyzing all the historical revisions of the software, the optimization method can calculate the quality of each module, based on the warnings issued by static analyzer as well as the mutual influence between modules. Warnings from the module with low quality are likely to correspond to real defects. The experiment results indicate that the optimization method can significantly reduce the false positive rate in static analysis. By investigating the effectiveness of privacy protection in Android, we that can help system developers to identify the vulnerabilities which can be exploited by malicious apps, and guide the users to protect their privacy data effectively.