TWC: Medium: Designing Strongly Obfuscated Hardware with Quantifiable Security against Reverse Engineering

University Of Massachusetts Amherst, Amherst MA

Investigators

Abstract

Our world has become increasingly reliant on integrated circuits (ICs). Mobile phones are deeply enmeshed in our everyday lives, we drive cars equipped with hundreds of ICs, and have come to depend on the power grid and other cyber physical systems that are controlled by ICs. Not surprisingly, the issue of securing hardware has become increasingly vital. A reverse engineering adversary may, for example, be motivated by extracting intellectual property from a circuit, cloning a design for product piracy, or creating a targeted backdoor for stealing cryptographic keys. In many real-world settings, we have to assume a strong attacker who has full access to the IC and can analyze all hardware features therein. Numerous hardware attacks over the last few years have shown that reverse engineering is a surprisingly easy task. The goal of this project is to develop strong and quantifiable protection mechanisms for ICs. This interdisciplinary project combines security engineering, formal verification and circuit design, with the aim of creating strong obfuscation techniques that thwart IC reverse engineering. The approach is based on integrated circuits that perform a substantially different function than the one an attacker would infer from its physical structure. The investigators will explore a broad design space of low-level stealthy circuit manipulations and higher-level obfuscation techniques that remove structural and functional boundaries. A crucial component of the project is the development of sound metrics to measure the level of achieved security. As part of community engagement, this project envisions a De-obfuscation Challenge that will involve research groups from academia and industry.