ISSN 1000-1239 CN 11-1777/TP

• 系统结构 •

### 电路工作模式下惰性节点的确定

1. (信息物理社会可信服务计算教育部重点实验室(重庆大学) 重庆 400044) (重庆大学计算机学院 重庆 400044) (xiaotong.sd@gmail.com)
• 出版日期: 2017-01-01
• 基金资助:
国家自然科学基金项目(61472052)；国家“八六三”高技术研究发展计划基金项目(2015AA015304,2013AA013202) This work was supported by the National Natural Science Foundation of China (61472052) and the National High Technology Research and Development Program of China (863 Program)(2015AA015304, 2013AA013202).

### Identifying Inactive Nets in Function Mode of Circuits

Cui Xiaotong, Zou Minhui, Wu Kaijie

1. (Key Laboratory of Dependable Service Computing in Cyber Physical Society (Chongqing University), Ministry of Education, Chongqing 400044) (College of Computer Science, Chongqing University, Chongqing 400044)
• Online: 2017-01-01

Abstract: The globalization trend of design and manufacture of IC raises serious concerns about hardware security since there is possibility that in each phase of design and manufacture hardware Trojan can be inserted. From the defender’s perspective, hardware Trojan in the host circuit may stay inactive for most of time but will result in disastrous consequences once activated, such as information leakage, false output, system crash, etc. As far as an attacker concerned, one of important design criteria is to prevent the Trojan circuit from being accidently activated. It is believed that inactive nets with lower switching probabilities in the circuit are more likely to be selected as the trigger signals of Trojan circuits. Hence finding these inactive nets is one of the existing countermeasures. However, current techniques only focus on finding inactive nets in test mode of the circuit. This paper proposes a method that can find inactive nets in function mode of the testing circuit. From an attacker’s point of view, the nets that are inactive in both test mode and function mode are the best candidates for Trojan triggers—it will result in the lowest probability of accidental activation of Trojan circuits in both modes. Hence for a defender, focusing on these nets will improve the efficiency of Trojan detection significantly.