Enhancing security via provably trustworthy hardware intellectual property

We introduce a novel hardware intellectual property acquisition protocol, show how it can support the transfer of provably trustworthy modules between hardware IP producers and consumers, and discuss what it might mean for a device to be considered “secure.” Specifically, we demonstrate the applicability of previous work in the software field of Proof-Carrying Code (PCC) to the problem of hardware trust and use it to combat the threat of hardware IP-level Trojans. We outline a semantic model representing the constructs permissible in a Verilog hardware description language (HDL) and show how this model can be used to reason about the trustworthiness of circuits represented at the register-transfer level (RTL). A discussion of “security-related properties” reveals how rules for trustworthy operation might be established for a particular design without necessarily specifying exact functionality. We then examine a hypothetical scenario involving a consumer with certain security needs and show how our system could be employed to guarantee that these needs are met by a hardware IP vendor's code.