Investigations and Development in the Area of Automated Security Evaluation of Android Devices with Focus on Bluetooth

Sammanfattning: Bluetooth is a technology that has been implemented in over 5 billion devices and therefore has a considerable impact. It is the dominant technology for shortrange wireless communication. Modern society relies heavily on information technology (IT), and this has introduced a significant threat to society and companies in the form of hackers whether they be state-sponsored, political activists, or part of organized crime. This has introduced the need for companies and organizations that strive to make devices more secure, as well as standards that can be used for evaluating how secure a device is. Common Criteria (CC) is an internationally recognized set of guidelines and standards that can be used for security evaluation. There is a growing demand for enhanced efficiency in the field of security evaluation, especially considering the move to agile methodologies in information and communication technology (ICT) product development. Historically, security evaluation has been tailored to each individual product. The current trends in the certification and global ICT evaluation industry indicate a move in the direction of a greater reliance on predefined test cases. In this thesis, I describe how I designed, developed, and evaluated a toolkit that automates the evaluation of Android devices concerning a selection of security requirements that concern Bluetooth from the Mobile Device Fundamentals Protection Profile in CC. This involved a literature study, examination of the Bluetooth Core Specification, software development, and evaluation of the toolkit. My results from evaluating the toolkit found that it only reports non-compliance with a security requirement if the target of evaluation (TOE) is non-compliant. Additionally, every time the toolkit reported compliance with a security requirement, manual evaluation verified that the TOE truly complied with the security requirement. Finally, during the development phase, I discovered a vulnerability that had not been discovered during manual evaluation. It has been confirmed by the developer to be a vulnerability and a patch is currently being developed. My evaluation indicates that the toolkit I have developed is reliable and that it could therefore be used in the security industry. By finding a vulnerability by using automation, I have shown that automation could potentially be a useful approach for vulnerability research. Similarly to fuzzing, automation can be used to expose a system to behavior that it does not expect and therefore potentially reveal vulnerabilities.