CAREER: Fortifying Leaky Hardware Interfaces with Distinguishability Set Architectures

Computing on personal data is a double-edged sword. On one hand, it enables revolutionary new applications such as personalized medicine and disease prediction. On the other hand, it runs the risk of revealing said personal data to unwanted parties. For example, using personal data on today’s processor chips can reveal that data through traces that the processor leaves behind. To make matters worse, different processors leave behind different traces, revealing different information, depending on how they were designed. This project will develop techniques to prevent data leakage through processors, for existing and future processor chips. The technical approach is to design a Distinguishability Set Architecture (DSA) for existing and future processors. DSAs are peers to existing Instruction Set Architectures (ISAs). Whereas the ISA specifies the functionality of each instruction, the DSA specifies under what conditions each instruction reveals secret information. With a DSA, programmers or compilers can tune sensitive programs to avoid leaking secrets. The first project thrust will develop DSA foundations, answering questions such as what should a DSA look like and how to capture leakage through various processor optimizations. The second thrust will develop compilers and hardware that use DSAs to improve program security. By precisely describing when and how processors reveal secrets, DSAs will unlock innovation on both software and hardware fronts. On the software side, programmers can focus on applications while DSA-aware compilers translate those applications to secure variants fit to run on different processors. On the hardware side, architects can use DSAs to reason about the privacy implications of hardware optimizations. The project will train a new class of students and researchers who can work across formal specifications, micro-architecture and compilers to build secure systems and, in the future, apply the lessons learned to other privacy-related problems. The DSA project will store all publications, code, and data-sets on public-facing websites, hosted at the University of Illinois for at least 3 years after the end of the project. This information will be made available via commercial websites. Links to these websites will be mirrored at http://cwfletcher.net/dsa. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.