Automated Security Analysis of Firmware

Sammanfattning

Internet of Things (IoT) devices are omnipresent in both industries and our homes. These devices are controlled by a software called firmware. Firmware like any other software may contain security vulnerabilities and the sole entity responsible to provide secure firmware is the vendor of the device. There are many analysis reports for individual devices that have found vulnerabilities however this is a manual process and unscalable. Several tools exist today for security analysis but usually with focus on a single aspect of firmware analysis. In this thesis, to better understand challenges with respect to automation of securityfocused analysis of firmware images in largescale, a framework is implemented by combining a number of existing tools using both static and dynamic analysis. A dataset of 241 firmware images from 7 vendors was collected. After evaluation of the framework on the dataset, it was shown that 85 images contained several known vulnerabilities which for some images a high percentage of these already were known before the firmware release date.