Malicious Trigger Discovery in FPGA Firmware
Abstract
FPGA netlists typically consist of hundreds to thousands of decoding structures that perform device operations. Malicious circuits (Trojans) often use complex decoders to delay activation until a time of an adversarys choosing. Identifying these triggers in a sea of legitimate decodings is highly desirable, but difficult to achieve, even with the assistance of a subject matter expert. This paper presents automation, developed on the DARPA VET program, that quickly discovers suspicious trigger circuits and verifies the correct operation for legitimate decoders.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 12, 2018
- Accession Number
- AD1052297
Entities
People
- John Hallman
- Scott Harper
- Stephen Baka
- Timothy Dunham