Memory System Random Number Generation Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The randomness of entropy input in memory systems with nonvolatile memory, such as SSDs, decreases over time due to deterioration in the entropy source, leading to a decrease in the security strength of pseudo random numbers generated.
Innovation Solution
A memory system with a processor that dynamically adjusts the length of the random number bits generated by an entropy source based on calculated randomness values, ensuring that pseudo random numbers are generated with stable security strength by updating the threshold values and random number bit lengths accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed threshold value is used for randomness evaluation, then the system operation is simple, but the security strength of pseudo random numbers decreases over time due to entropy source deterioration
Solution Approach 1:
The patent implements dynamic threshold values that automatically adjust based on the age of the entropy source. The threshold increases as the entropy source ages, compensating for expected deterioration and maintaining consistent security strength over time. This transforms a static evaluation system into a dynamic one that adapts to temporal changes in entropy quality.
Solution Approach 2:
The patent changes the threshold parameter based on the operational age of the entropy source. By adjusting the threshold value as a function of time/age, the system maintains appropriate security requirements despite the natural deterioration of the entropy source, balancing security needs with system complexity.
2Reliability
If the randomness evaluation is performed frequently, then the security strength is maintained, but the processing time and system overhead increase
Solution Approach 1:
The patent implements periodic randomness evaluation at predetermined time intervals rather than continuous evaluation. This periodic approach maintains security by regularly checking entropy quality while minimizing processing overhead by avoiding unnecessary frequent evaluations between check points.
Solution Approach 2:
The patent performs randomness evaluation in advance before pseudo random number generation is required. This preliminary evaluation ensures that entropy quality is verified beforehand, preventing security issues while allowing efficient batch processing and reducing time-critical overhead during actual random number generation operations.
Data Source
AI summary
A memory system connectable to a host device includes a nonvolatile memory, a first circuit configured to generate a sequence of random number bits, and a processor configured to instruct the circuit to generate a sequence of random number bits having a first length, calculate a first value indicating randomness of the sequence, determine whether the first value exceeds a threshold value, upon determining that the first value exceeds the threshold value, generate a pseudo random number using the sequence, upon determining that the first value does not exceed the threshold value, instruct the first circuit to generate another sequence of random number bits having a second length greater than the first length, and generate a pseudo random number using said another sequence, and write or read data to or from the nonvolatile memory using the generated pseudo random number.