Memristive Hash Function Using Write Disturb Entropy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current secure hash functions, particularly those based on mathematical algorithms, fail to provide sufficient performance, power, and area efficiency, and lack robustness in hardware implementations, especially when used for cryptographic applications like digital signatures and integrity validation.
Innovation Solution
A memristive hardware hash function that utilizes discrete quasi-stable states and differential reads in a memristor crossbar array to increase entropy and robustness, leveraging manufacturing variations and the write disturb phenomenon to create a unique and unclonable key for each instance, effectively forming a keyed-hash message authentication code without separate key management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If pure software implementations of hash algorithms are used, then flexibility and ease of implementation are improved, but performance and security are insufficient
Solution Approach 1:
The patent replaces software-based hash algorithms with a hardware implementation using memristive devices. The mechanical/computational system of software execution is substituted with a physical system that leverages intrinsic properties of memristors (non-linearity, non-volatility, sensitivity to process variations) to perform hash functions directly in hardware, thereby improving both performance and security simultaneously
2Productivity
If hardware accelerators using digital logic are used, then performance is improved, but power consumption and area are increased
Solution Approach 1:
The patent changes the fundamental parameters of the computing system by transitioning from digital logic circuits to analog/memristive devices. This parameter change enables the system to achieve high performance through parallel analog computations and non-volatile memory properties, while consuming significantly less power due to the inherent low-power characteristics of memristive devices and elimination of continuous clocking required by digital logic
3Productivity
If hardware accelerators using digital logic are used, then performance is improved, but device complexity and area are increased
Solution Approach 1:
The patent creates a universal memristive platform that can perform multiple cryptographic functions (hash algorithms, digital signatures, integrity validation) using the same hardware architecture. The memristive crossbar array serves as a multi-functional device that can implement different hash functions and cryptographic primitives, thereby reducing overall system complexity compared to dedicated digital logic circuits for each function
4Use of energy by moving object
If intrinsic properties of hardware are used to create entropy, then power and area efficiency are improved, but robustness is insufficient
Solution Approach 1:
The patent implements feedback mechanisms in the memristive hash function, where the output of each processing stage is fed back into subsequent stages. This feedback structure, combined with the non-linear dynamics of memristive devices, enhances robustness by creating complex, unpredictable behavior that is sensitive to initial conditions but deterministic in operation, thereby improving reliability while maintaining power efficiency
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
The memristive hash function achieves low power and area efficiency while providing robustness and high entropy, suitable for applications requiring integrity checks and secure key generation, with statistical characteristics demonstrating close-to-optimal uniqueness and diffuseness, effectively resisting modeling attacks and brute-force attempts.
Implementation Method 1
The array may be initiated into a state in which changes to the neighboring cells are predictable according to a respective proximity to a currently selected cell
Implementation Method 2
leveraging manufacturing variation in the memristor cells to increase the entropy of the hash
Implementation Method 3
A function according to the present embodiments may use discrete quasi-stable states, in which the memristor drift rates are sufficiently slow
Data Source
AI summary
Apparatus and method for hashing a message, comprises using an array of individually selectable memristor cells. The memristor cells are subject to write disturb that affects cells neighboring a selected cell so that a write operation into one cell has a knock-on effect on the neighbors. The array is initiated into a known stable state so that these changes to neighboring cells are predictable according to proximity to the currently selected cell. An inserter sequentially mixes bits with the hash so far to insert bits into successively selected cells of the memristor array and forms a succession of memristor array states including the knock on effects on the neighboring cells. A final resulting memristor array state following input of the bits forms the hash of the message.