Research On Suspicious Circuit Structure Analysis Based Hardwear Trojan Detection Technology

The security of integrated circuit(IC) gained widespread attention in academia, government agencies, and manufacturing currently. One part of the reason is the design and manufacturing outsourcing migration to low-cost regions worldwide, another part is the increasing dependence from different suppliers of design automation tools and the third-party intellectual property(IP), such that the current supply chain of an integrated circuit is considered more than ever vulnerable to malicious modified, so as soon as possible to carry out the reliability of chip design and hardware Trojans research without delay.This thesis based on the basic characteristics of hardware Trojans design presented a Trojan detection technique based on the analysis of suspicious circuit structure, This detection technique can be applied to detect hardware Trojans on gate-level netlist of any IC design stage, not only to ensure the credibility of the chip design phase, it can also determine the phase which a Trojan was inserted in. the study proved that the detection method can effectively locate the distinctive characteristics present in the circuit or high hidden hardware Trojans, for the area very small and concealed hardware Trojan can achieve very good detection effect, can effectively detect an area of not less than 0.0038% hidden Trojan. The main research work of this paper includes the following:(1) Hardware Trojans characteristics and study of existing detection methods. Analysis of the essential characteristics of hardware Trojans and principles, summarize the advantages and disadvantages of existing detection methods, combined with existing detection technology from hardware Trojan hidden qualities proposed the hardware Trojan detection technology based on suspicious circuit structure analysis.(2) Analysis of suspicious circuit structure. Starting from the basic characteristics of hardware Trojans, summed up the basic circuit structure which is most likely a hardware Trojan. Using the powerful text processing capabilities and regular pattern matching of Perl language to turn the under test circuit netlist into the intermediate data storage form of easy to structure traversal. Research using Perl traversal matching each suspicious circuit structure, focusing on analysis and comparison of two different methods to obtain low activity suspicious circuit node through a large number of data analysis and processing to determine the applicability of the two methods.(3) Design and detection research of hardware Trojans. In the RTL and gate-level of UART reference circuit, this thesis designs five different types of structure hardware Trojans whose size ranging from 1.05% to 71.52%, and programming of different Trojans circuit random into a different location of the original circuit netlist. Achieve detection and location Trojan horse circuit in the 10 under test circuit of UART circuit that insertion designed Trojan horse and analyze the Trojans circuit functions.(4) Building automation platform and comparing test results. Use Perl::Tk module to build automated testing platform, achieve intuitive and efficient hardware Trojans detection on infected AES circuits. In the original circuit AES randomly inserted Trojan horses whose size ranging from 0.0014% to 0.026%, then started the hardware Trojan detection, compared the detection results with other literature, the results show that the detection method in this thesis detects the small horse better.