Grant: Discovery Grants Program - Individual

With widespread growth of Internet of Things (IoT) and cyber-physical systems, our reliance on digital systems has been dramatically increased. However, this growing popularity also turns IoT into a common target for cybersecurity threats. Specifically, connectivity to the Internet and heterogeneity by design coupled with shorter time to market potentially cause significant increases in vulnerabilities in an IoT-based system. As evidenced from recent cyberattacks (e.g., Industroyer and SolarWinds), such vulnerabilities are frequently exploited to create nationwide (and even worldwide) catastrophic effects (including failure of critical infrastructure). Many of those security vulnerabilities result from flaws in the underlying device software/firmware. Therefore, hunting vulnerabilities (which includes both identifying and understanding them) in IoT firmware is essential to reduce the chances of possible cyberthreats and mitigate their impacts in the future. However, in the absence of source code for most proprietary IoT devices, vulnerability hunting from binary code becomes extremely challenging. Among many, notable challenges include diverse usage of hardware architectures, software and code transformations, growing scale of connected IoT devices and their binary code, and missing IoT domain-specific datasets for binary code analysis. In this proposed research program, we will investigate and address those major challenges and build a hybrid binary code analysis framework to accurately identify security vulnerabilities in large-scale IoT firmware. To that end, this program will aim at first building an IoT domain-specific vulnerability and firmware databases, then proposing a static analysis based technique to detect, locate, remediate known vulnerabilities in an IoT firmware, and finally devising scalable dynamic solutions that validate known vulnerabilities as well as detect unknown vulnerabilities in IoT device firmware. The proposed program will have significant impact in several critical areas:enabling vulnerability hunting in IoT firmware through binary code analysis with the potential of applying the proposed algorithms, technologies and tools to other cyber-physical systems; publishing vulnerability and firmware databases for IoT devices to help future research (beyond ours); and training highly qualified personnel with research, professional and technical skills on the demanding fields of IoT security and vulnerability assessment. In the long term, this program will develop adaptive solutions for vulnerability hunting for other cyber-physical systems and addressing novel challenges, e.g., correcting identified vulnerabilities in near real-time and measuring security impacts of vulnerabilities. Thus, the proposed program is designed to continuously contribute innovations in cybersecurity for emerging technologies as well as the next generation of skilled workforce to meet the ever-growing needs of the Canadian cybersecurity industry.