Hardware Trojan detection by specifying malicious circuit properties

Abstract

Ahstract-This work addresses the increasing danger of ship ping integrated circuits (either fully digital or mixed signal) that contain malicious hardware modifications. Potential security threads are established by so called hardware Trojans, imple mented in the physical silicon structure. A desire of system engineers is to identify such back-door functionalities during an early design phase. The paper discusses how techniques deduced from model checking applications can be used for Trojan detection. A set of potential malicious properties is defined and handed to an automatic tool. Any returned counterexample identifies a feasible attack path and is a basis for further detailed inspection. *

Towards a Hardware Trojan Detection Cycle

“…Many of the examinations and checks can be automated. It therefore complements approaches like [2] and partly incorporates [8]. Rule sets and knowledge base can be shared among industry, similar to personal computer anti-virus companies sharing their knowledge and fingerprints with each other.…”

ForASec: Formal Analysis of Hardware Trojan-Based Security Vulnerabilities in Sequential Circuits

“…To address these research challenges, we propose a novel and generic framework for formal analysis of security vulnerabilities (ForASec, Section 4) to perform the security analysis while considering the complete coverage of parametric behavior (i.e., leakage power, dynamic power and propagation delay) and process variations. The proposed ForASec consists of the following key components: [20], [21], [22], [23] [26], [27] [24]…”

“…Typically, model based simulations are used for analyzing the security vulnerabilities [14], [15], but they cannot cover all possible test cases in complex systems because of their computational constraints (energy and memory) [16] and floating point inaccuracies [17]. To ensure the completeness and accuracy, mathematical modeling and formal verification based vulnerability analysis techniques have been proposed [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], as shown in Table 1. Although, to some extent, mathematical modeling can overcome above-mentioned limitations, it is still prone to human error and increases the design time.…”

“…On the other hand, functional [20], [21], [22] and behavioral [23] model checking based approaches provide the comprehensive vulnerability analysis and cover all possible input test cases but they inherently pose the state-space explosion problem for complex systems. To overcome this issue, several approaches for performance modeling have been proposed, which model the side-channel parameters, i.e., power [24], [25], [26], propagation delay [25], [26] and temperature [27], and analyze the vulnerabilities based on temporal properties [25], [26], [27] and property specification language [24].…”