IC Failure Detection via Boolean Derivative Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Integrated circuits (ICs) are vulnerable to physical attacks due to Single Points Of Failure (SPOFs), which can be exploited by attackers to compromise security, and existing detection methods are inadequate in identifying and mitigating these vulnerabilities, especially in the presence of Focused Ion Beam (FIB) attacks.
Innovation Solution
A device and method for detecting Points Of Failure in ICs by iteratively selecting n-tuples of elements in the structural description, modifying them to test if the derivative of the Boolean function associated with sensitive functional blocks is zero, and identifying SPOFs or Multiple Points Of Failure (MPOFs) based on this condition, allowing for remediation and enhanced security measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If alarm signals are diversified using simple duplication or multiplication, then security against single point failures is improved, but device complexity increases and manual implementation may result in missed signals or synthesis optimization that reintroduces vulnerabilities
Solution Approach 1:
The alarm signal is segmented into multiple independent alarm signals, each derived from different subsets of the alarm condition. This segmentation ensures that a single point failure cannot affect all alarm signals simultaneously, while the automated generation process manages the complexity of creating and maintaining these segmented signals.
Solution Approach 2:
The system performs preliminary action by automatically generating diversified alarm signals during the design phase using ATPG methods. This preliminary automation prevents manual implementation errors and ensures that the diversification is properly established before the circuit is manufactured, avoiding the need for complex manual wiring and reducing the risk of missed signals.
2Difficulty of detecting and measuring
If ATPG-based detection methods are used to identify single points of failure, then detection capability is improved, but the method requires complex test vector generation and state data analysis
Solution Approach 1:
The system creates copies of test vectors and state data through automated pattern generation. Instead of manually creating complex test scenarios, the ATPG methodology automatically generates multiple copies of test vectors that can be systematically applied to detect single points of failure, reducing the complexity of the detection process while maintaining high detection capability.
Solution Approach 2:
The ATPG-based detection method incorporates feedback loops where test results are analyzed and used to generate improved test vectors. This feedback mechanism automatically refines the detection process, identifying single points of failure more efficiently without requiring manual intervention to increase complexity, as the system learns and adapts from each test iteration.
3Object-affected harmful factors
If FIB attacks are used to cut vulnerable wires, then attackers can cancel alarms and access secrets, but FIB editing is difficult due to routing congestion and tool reliability issues
Solution Approach 1:
The system applies local quality by creating alarm signals with different local characteristics - each alarm signal is derived from different local subsets of the alarm condition and different local wires. This ensures that even if an attacker can access certain local areas with FIB, they cannot easily cancel all alarm signals, as each alarm has its own local quality and dependency on different circuit elements.
Solution Approach 2:
The invention transitions from a single-dimension alarm signal to a multi-dimensional alarm system where alarms are diversified across multiple dimensions - different signal paths, different logical derivations, and different physical locations. This dimensional diversification makes FIB attacks more difficult because the attacker would need to successfully edit multiple independent elements across different dimensions of the circuit rather than a single vulnerable wire.
4Manufacturing precision
If post-production foundry tests are used to verify signal continuity, then manufacturing quality is improved, but these tests do not account for the pertinence of tested signals to security vulnerabilities
Solution Approach 1:
The ATPG-based detection system acts as an intermediary between standard manufacturing tests and security vulnerability assessment. It takes the signal continuity information from post-production tests and adds the intermediary layer of security-relevant analysis, determining which signals are pertinent to security vulnerabilities. This intermediary process bridges the gap between general manufacturing quality control and specific security vulnerability detection.
Data Source
AI summary
Devices, methods, and computer program products for detecting Points Of Failures in an integrated circuit (IC) are provided. The integrated circuit device is described by a structural description (2) comprising a plurality of elements, the elements representing cells and wires interconnecting the cells, the structural description further comprising portions representing a set of sensitive functional blocks (16), each sensitive functional block comprising one or more inputs, at least one sensitive output, and a set of elements interconnected such that the value of the sensitive output is a Boolean function of the input values of the sensitive functional block. The detection device (100) comprises:a selection unit (101) configured to iteratively select a n-tuple of elements in at least the portions of the netlist corresponding to said sensitive functional blocks,a testing unit (104) configured to test each selected n-tuple of elements, the testing unit being configured to:modify said selected n-tuple of elements from an initial state to a testing state;determine if the derivative of the Boolean function associated with each sensitive functional block is equal to zero.The detection device (100) is configured to detect that said n-tuple represents a Point Of Failure of order n in the integrated circuit (IC) device if the derivative of the Boolean function associated with said sensitive functional block is equal to zero.


