Memristive Random Bit Generation for Hardware-Unique Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current hardware-based truly random number generators are either slow or require significant hardware, while software-based pseudo random number generators are reproducible and cannot be tied to unique hardware, failing to provide a persistent, passive means for unique hardware identification.
Innovation Solution
The use of memristive devices with a lateral switching process, where two memristive devices are set to a low resistance state and then reset laterally, resulting in a truly random binary bit due to non-uniform material variations, allowing for persistent random number generation and unique hardware identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current hardware-based truly random number generators use thermal noise in a resistor or avalanche noise from an avalanche diode, then truly random numbers are generated, but the generation speed is slow or significant hardware is required
Solution Approach 1:
The patent changes the physical parameter being exploited from thermal noise or avalanche noise to resistance variations in memristive devices. By switching to a different physical mechanism (resistance state variations instead of noise), the system achieves both true randomness and faster generation speeds with reduced hardware requirements
Solution Approach 2:
The patent replaces traditional noise-based physical mechanisms (thermal noise, avalanche noise) with a resistance-based mechanism in memristive devices. This substitution allows for faster response times and lower hardware complexity while maintaining true randomness through inherent material variations
2Productivity
If software-based pseudo random number generators are used, then large series are generated quickly, but the generation is reproducible and cannot be tied to unique hardware
Solution Approach 1:
The memristive devices serve themselves by utilizing inherent material variations that are unique to each device. The random number generation capability is embedded in the physical structure of the device itself, eliminating the need for external seeding mechanisms while ensuring hardware-specific uniqueness
Solution Approach 2:
The patent extracts the random number generation capability from software algorithms and embeds it directly in the hardware structure of memristive devices. By taking out the dependency on software seeding and placing the randomness source in the physical device characteristics, the system achieves both speed and hardware uniqueness
3Reliability
If inherent variations in active device components are used for physically unclonable functions, then unique hardware identification is achieved, but energy overhead increases
Solution Approach 1:
The memristive devices serve multiple functions: they generate truly random numbers and simultaneously provide unique hardware identification through their inherent variations. This multi-functionality eliminates the need for separate PUF circuits, reducing overall energy overhead while maintaining hardware uniqueness
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method generates a truly random binary bit that can be retained in a passive medium for future queries, enabling unique hardware identification and overcoming the limitations of existing technologies.
Implementation Method 1
A typical memristive device has two critical resistance states, a high resistance state (HRS) and a low resistance state (LRS). To change a memristive device from the HRS to the LRS (termed a SET operation), a voltage bias of the appropriate polarity and magnitude, VSET, must be applied to the device.
Implementation Method 2
Once the device is in the LRS, it may be returned to the HRS via a RESET operation, typically by applying a lower voltage, VRESET, but without the current limit.
Implementation Method 3
This invention generates a truly random binary bit, where the resistance of memristive devices serve as the analogue bit values. Said resistance values and the randomness by which they are generated stems from the inherently non-uniform, irreproducible variations in the materials of which the system is composed.
Data Source
AI summary
Apparatus for generating a truly random binary bit value in which the resistance of memristive devices serve as the analogue bit values. Resistance values and the randomness by which they are generated stems from the inherently non-uniform, irreproducible variations in the materials of which the system is composed.