Creating A Graph Dataset For The Android Framework

Context

Static- and Dynamic Taint Analysis is a technique to track the flow of information in software. However, the analysis of the Android Framework is challenging because of the complexity of the framework. Therefore, it is essential to create a graph dataset of the Android Framework to analyze the flow of sensitive information in the framework. Currently, there is no dataset available that represents the Android Framework as a graph. Aim of this project is to create a graph dataset of the Android Framework that represents the code in a structure that can effectively be queried for taint analysis.

Motivation

We want to analyze the flow of sensitive information in the Android Framework. To achieve this goal, we need a graph dataset that represents the Android Framework. There exists several tools that can generate and visualize graphs from source code. The student will learn in this project to use these tools and generate a graph dataset of the Android Framework.

Goal

What do we want to achieve?

• An effective method to create a graph dataset of the Android Framework that can be used for taint tracking.

What do we want the student to deliver?

• An analysis of existing graph generation tools.
• A graph dataset of the Android Framework.

Requirements

What do we want the student to bring to the project?

• Very good skills in programming Java, and Python
• Basic knowledge about Android
• Interest in Taint Analysis
• Interest in Graphs to Code representations
• Interest in Software Engineering
• Interest in Software Analysis

Pointers

What resources and other related work could help the student to work on this project?

•  Graph Databases
•  Flowdroid
•  Joern
•  Code Property Graph
•  Frauenhofer: Code Property Graph
•  Taint Analysis Lecture