Hardware Trojan detection and localization based on local detectors

Authors: AMIN BAZZAZI, MOHAMMAD TAGHI MANZURI SHALMANI, ALI MOHAMMAD AFSHIN HEMMATYAR

Abstract: Hardware Trojans are one of the serious threats with detrimental, irreparable effects on the functionality, security, and performance of digital integrated circuits. It is difficult to detect Trojans because of their diversity in size and performance. While the majority of current methods focus on Trojan detection during chip testing, run-time techniques can be employed to gain unique advantages. This paper proposes a method based on the online scalable detection technique, which eliminates the need for a reference chip. Involving local detectors, this technique assesses the variations in the logical values of each node to find out whether there are Trojans. This method excludes time and power measurements, which are common parameters in most conventional methods. The detectors provide Trojan-localization capability in our proposed technique. Two remarkable features of this technique are low power and low area overhead. The results are reported by simulation and implementation of common benchmarks, which show the high Trojan detection rate of the proposed method.

Keywords: Hardware Trojans, run-time method, local detectors, detection, localization

Full Text: PDF