Self-Timed Ring Oscillator RNG for High-Entropy Fast Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing random number generators struggle to produce high-speed, unpredictable random numbers with high entropy, which is crucial for secure systems but often suffer from periodicity and bias in output values.
Innovation Solution
The implementation of a self-timed ring (STR) oscillator-based random number generator that includes an initial random number generator, a duty corrector, and a sampling circuit to generate and correct output values, ensuring a 50% duty ratio and random token positioning for each sampling clock, thereby enhancing entropy and unpredictability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a traditional random number generator uses thermal noise or shot noise to generate random numbers, then the randomness is improved, but the generation speed is limited and periodicity may occur
Solution Approach 1:
The patent replaces traditional thermal/shot noise-based random number generation with a self-timed ring oscillator system that uses digital logic circuits and clock signals. This substitution of physical noise mechanisms with a structured electronic oscillation system enables faster generation speeds while maintaining randomness through duty cycle variations and token/bubble propagation in the ring stages.
Solution Approach 2:
The patent introduces dynamic elements by using a self-timed ring oscillator where the oscillation period and duty cycle vary randomly. The ring stages dynamically propagate tokens and bubbles based on random duty cycle values, creating unpredictable output sequences that maintain high randomness while achieving faster generation speeds through the oscillatory mechanism.
2Productivity
If a random number generator operates at high speed, then the productivity is improved, but the entropy and unpredictability of the generated numbers may decrease
Solution Approach 1:
The patent implements feedback mechanisms where the ring oscillator's output is fed back through multiple ring stages, and the duty corrector continuously adjusts duty cycles based on previous output states. This feedback ensures that even at high operating speeds, the system maintains entropy by preventing periodic patterns and ensuring unpredictable token/bubble propagation through the ring stages.
Solution Approach 2:
The patent changes key parameters dynamically - specifically the duty cycle of clock signals and the timing of token/bubble propagation through ring stages. By randomly varying these temporal parameters at each oscillation cycle, the system maintains high entropy and unpredictability in the generated random numbers while operating at high speeds.
3Reliability
If a random number generator uses complex circuitry to ensure high entropy, then the reliability is improved, but the device complexity increases
Solution Approach 1:
The patent achieves multi-functionality where the self-timed ring oscillator serves multiple purposes: generating clock signals, creating random duty cycle variations, propagating tokens and bubbles, and producing the final random number output. This universal use of the ring oscillator structure reduces overall device complexity while maintaining high entropy quality through its inherent oscillatory and randomizing properties.
Solution Approach 2:
The patent segments the random number generation process into distinct functional blocks: an initial random number generator for seeding, a self-timed ring oscillator for core generation, a duty corrector for adjustment, and a sampling circuit for final output. This segmentation allows each component to be optimized independently, reducing overall complexity while maintaining high entropy through the coordinated operation of specialized subsystems.
4Reliability
If a random number generator uses thermal noise or shot noise, then the true randomness is achieved, but the output may exhibit periodicity and bias
Solution Approach 1:
The patent applies preliminary action by using an initial random number generator to seed the self-timed ring oscillator before main operation begins. This pre-initialization ensures that the ring oscillator starts with high-entropy seed values, preventing periodicity and bias from the outset. The duty corrector also performs preliminary adjustment of duty cycles to ensure uniform 50% distribution before random number generation proceeds.
Solution Approach 2:
The patent uses copying by propagating tokens and bubbles through multiple ring stages, where each stage copies and transforms the random state from previous stages. This copying mechanism through the ring oscillator chain amplifies and distributes the initial random entropy while the duty corrector ensures each copied signal maintains uniform 50% duty cycle, eliminating bias in the final output.
Data Source
AI summary
A random number generator according to example embodiments includes an initial random number generator configured to generate an initial random number, a self-timed ring (STR) oscillator configured to receive the initial random number, the STR oscillator having a plurality of ring stages generating, in response to a clock, either a bubble that does not change an output state of a previous clock or a token changing the output state of the previous clock, a duty corrector configured to adjust a duty of each of output values of the ring stages, and a sampling circuit configured to sample a random number using a logic operation from the duty-corrected output values.


