PRNG Seed Generation via Entropy Sampling and Hashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating seed values for pseudo random number generators (PRNGs) are insufficient in ensuring independence, non-repeatability, and numerical dispersion, particularly in distributed computing environments, leading to potential weaknesses and predictability.
Innovation Solution
A technique that generates seed values independent of time, location, and hardware attributes, using multiple sampling and whitening functions to produce uniformly distributed and unique seeds across a wide range of input values, suitable for various computing hardware and operating systems, including those with synchronized clocks and virtualized environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If seeds are derived from host computer calendar and clock values, then seed generation is simple, but independence and non-repeatability are insufficient
Solution Approach 1:
The patent segments the seed generation process into multiple independent components: collecting entropy from multiple sources (clock values, hardware attributes, process identifiers), processing each through separate cryptographic hash functions, and combining the results. This segmentation ensures that no single predictable source can compromise the overall seed quality, resolving the contradiction between simplicity and reliability.
Solution Approach 2:
The patent creates a composite seed value by combining multiple entropy sources using cryptographic operations. The final seed is a composite of hardware attributes, temporal data, and process characteristics, processed through hash functions. This composite approach maintains operational simplicity while dramatically improving independence and non-repeatability compared to single-source seeding.
2Ease of manufacture
If seeds are derived from hardware attributes, then generation is straightforward, but numerical dispersion is insufficient
Solution Approach 1:
The patent transforms hardware attributes through cryptographic hash functions, fundamentally changing the parameter space. Instead of using raw hardware values directly, the system processes them through multiple hash iterations and combines them with other entropy sources, expanding the numerical range and improving dispersion across the seed space while maintaining straightforward implementation.
3Ease of operation
If PRNG executes on distributed processes with synchronized clocks, then coordination is simplified, but seed independence is compromised
Solution Approach 1:
The patent introduces cryptographic hash functions as intermediaries between the synchronized clock values and the final seed. Even when multiple distributed processes have identical or synchronized timestamps, the hash function processing combined with unique hardware attributes and process identifiers ensures that each process generates independent seeds. This intermediary transformation maintains coordination simplicity while preserving seed independence.
Data Source
AI summary
Two or more numerical samples are read from a storage location (or many storage locations) and the samples are compared to generate a single bit of entropy. This method is repeated to populate two arrays which are also compared to generate a single bit. Comparison of the arrays is also repeated to generate a sequence of bits which are formed into computer words. The words are combined with a shift register to generate a seed block array of any length. The seed block array is used to seed a pseudo random number generator. The storage location is a register, accumulator, buffer, clock, address, memory location, etc., that changes periodically. Comparison may be performed by counting the number of certain types of bits in the two values and returning a one or zero depending upon the count.