Random Number Generator Fault Injection Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Random number generators are vulnerable to fault injection attacks, which can distort their non-deterministic character and compromise security algorithms, as existing detection methods fail to detect attacks that alter the bit stream downstream of the noise source or within the ciphering algorithm.
Innovation Solution
A method involving interposing a bit pattern into the bit stream from a noise source, sampling downstream to detect the pattern's presence, and using a resilient function to generate a random number, thereby protecting against fault injection attacks by ensuring the pattern's integrity and randomness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If statistical tests are performed over variable depth to verify non-deterministic character, then detection capability improves, but device complexity increases
Solution Approach 1:
A known bit pattern is inserted into the random bit stream before processing. This preliminary action creates a detectable signature that can verify the integrity of the random number generation process without requiring complex statistical analysis during operation.
Solution Approach 2:
Instead of performing complex statistical tests on the entire random bit stream, the system copies a known reference pattern into the stream and simply checks for its presence. This replaces complex measurement with a simple detection operation.
2Reliability
If fault injection detection is implemented, then reliability improves, but ease of operation deteriorates
Solution Approach 1:
The random number generator performs self-verification by checking for the presence of its own inserted pattern. This self-service mechanism detects faults without requiring external verification systems, maintaining reliability while simplifying operation.
3Measurement precision
If bit pattern is interposed in bit stream, then fault injection detection capability improves, but manufacturing precision requirements increase
Solution Approach 1:
The detection mechanism is segmented into independent components: pattern generation, pattern insertion at predetermined positions, and pattern detection. This segmentation allows each component to be implemented with standard precision requirements rather than requiring high overall system precision.
Data Source
AI summary
A method for detecting a fault injection in a random number generation circuit, wherein a bit pattern is mixed to a bit stream originating from a noise source and the presence of this pattern is detected in a signal sampled downstream of the mix.