IP Trust: The Problem and Design/Validation-Based Solution

A rapidly growing third-party intellectual property (IP) market provides IP consumers with more options for designing electronic systems. It also reduces the development time and expertise needed to compete in a market where profit windows are very narrow. However, one key issue that has been neglected is the security of electronic systems built upon third-party IP cores. Historically, IP consumers have focused on IP functionality and performance than security. The prejudice against the development of robust security policies is reflected in the IP design flow (see Fig. 4.1), where IP core specification usually includes functionality and performance measurements. This lack of security assurance on third-party IPs is a major threat for the semiconductor industry. For example, a large number of side-channel-based attacks have been reported, which extract sensitive information from systems that were purportedly mathematically unbreakable [1–7]. The emergence of hardware Trojans (malicious logic) embedded in third-party IP cores has largely re-shaped the IP transaction market and there are currently no comprehensive detection schemes for identifying these Trojans. Some Trojan detection methods such as side-channel fingerprinting combined with statistical analysis have shown promising results, but most of the post-silicon stage Trojan detection methods rely on golden models which may not be available given the existence of untrusted IP cores.