Research On Hardware Trojan Detection And Recovery Mechanism And Cost Optimization Of Untrusted Third-Party IP Cores

The globalization process of the integrated circuit industry has aroused researchers' attention to Hardware Trojan horses.Due to the concealment and diversity of Hardware Trojans,detecting them during testing may consume a lot of resources.When the device under test contains a system-on-chip with an untrusted third-party intellectual property core(IP core),whether it is a system purchase cost issue or a security issue,it faces huge challenges.To this end,this thesis uses untrusted third-party IP cores as a threat model,based on Hardware Trojans without memory elements,and compares the logic errors and traditional errors caused by Trojan horse circuits,and proposes a security strategy for Trojan horse circuit detection and recovery.Based on the security strategy,with the goal of optimizing the cost of system procurement,the integer linear programming model solves the shortcomings of large-scale circuits,and the system procurement cost is optimized from the perspective of the graph theory model.And on this basis,explore the impact of the system time and area overhead on the experiment,and verify the impact of the threat model on the system cost overhead.Finally,design and build an online graph theoretical calculation platform,real-time online calculation of system costs,and output visual experimental results graphs.The main work of this thesis includes:First,a security strategy for the detection and recovery of Trojan horse circuits is proposed.Firstly,the third-party IP core is untrusted as the threat model of this thesis.Secondly,in view of the similarities between logic errors caused by Trojan horse circuits and traditional processor errors,a calculation framework for error detection and error recovery will be built based on space and time redundancy mechanisms.Finally,in view of the difference between the Trojan horse circuit and the traditional error and the existence of the Trojan horse circuit in the IP core,using the diversity of the IP core,the corresponding Trojan horse circuit detection and recovery security strategy will be formulated.Second,based on the security strategy and threat model,a task scheduling algorithm for system cost optimization is proposed.That is,graph theory system cost optimization model.On the one hand,analyze the existing integer linear programming model.On the other hand,a graph theory model is proposed.In the graph theory model,the task nodes in specific applications and their characteristics are analyzed,and ordinary application models are transformed into intuitive graph models.Finally,according to the security strategy of the Trojan horse circuit detection and recovery,the basic task scheduling mechanism,and time/area constraints,a resource allocation mechanism based on system procurement costs and time/area is established.Third,with online technology application as the core idea,we design and implement a system-on-chip Hardware Trojan horse detection graph theoretical calculation platform based on security strategies and graph theory cost optimization algorithms.The algorithms in abstract theoretical research are real-time calculation operators,and online experimental simulations are performed.And the online calculation experiment results are output in the visual chart.