AbstractIn-cabin connectivity and its enabling technologies have increased dramatically in recent years. Security was not considered an essential property, a mind-set that has shifted significantly due to the appearance of demonstrated vulnerabilities in these connected vehicles. Connectivity allows the possibility that an external attacker may compromise the security - and therefore the safety - of the vehicle. Many exploits have already been demonstrated in literature. One of the most pervasive connective technologies is
Bluetooth, a short-range wireless communication technology. Security issues with this technology are well-documented, albeit in other domains. A threat intelligence study was carried out to substantiate this motivation and finds that while the general trend is towards increasing (relative) security in automotive Bluetooth implementations, there is still significant technological lag
when compared to more traditional computing systems.
The main contribution of this thesis is a framework for the systematic security evaluation of the automotive Bluetooth interface from a black-box perspective (as technical specifications were loose or absent). Tests were performed through both the vehicle’s native connection and through Bluetooth-enabled aftermarket devices attached to the vehicle. This framework is supported through the use of attack trees and principles as outlined in the Penetration Testing Execution Standard. Furthermore, a proof-of-concept tool was developed to implement this framework in a semi-automated manner, to carry out testing on real-world vehicles. The tool also allows for severity classification of the results acquired, as outlined in the SAE J3061 Cybersecurity Guidebook for Cyber-Physical Vehicle Systems. Results of the severity classification are validated through domain expert review. Finally, how formal methods could be integrated into the framework and tool to improve confidence and rigour, and to demonstrate how future iterations of design could be improved is also explored.
In conclusion, there is a need for systematic security testing, based on the findings of the threat intelligence study. The systematic evaluation and the developed tool successfully found weaknesses in both the automotive Bluetooth interface and in the vehicle itself through Bluetooth-enabled aftermarket devices. Furthermore, the results of applying this framework provide a focus for counter-measure development and could be used as evidence in a security assurance case. The systematic evaluation framework also allows for formal methods to be introduced for added rigour and confidence. Demonstrations of how this might be performed (with case studies) were presented. Future recommendations include using this framework with more test vehicles and expanding on the existing attack trees that form the heart of the evaluation. Further work on the tool chain would also be desirable. This would enable further accuracy of any testing or modelling required, and would also take automation of the entire process further.
|Date of Award||2018|
|Supervisor||Siraj Shaikh (Supervisor), Olivier Haas (Supervisor), Alistair Ruddle (Supervisor) & Jeremy Bryans (Supervisor)|