Ferroelectric 2T-2C Memory Random Number Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing physical unclonable function (PUF) devices struggle with achieving high repeatability of random number generation, with about 15% of bits changing unpredictably between instantiations, which affects the security and predictability of the digital signature.
Innovation Solution
A method using a block of ferroelectric memory cells with a 2T-2C configuration, involving a 1T-1C write mode to set both sub-cells to the same data state, followed by a 2T-2C read and restore operation, and an imprinting process to enhance the repeatability of the random number generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional PUF approaches are used to generate random numbers, then the digital signature is unique to each device, but about 15% of bits change unpredictably between instantiations, reducing repeatability
Solution Approach 1:
The patent applies preliminary action by performing a write operation to a first memory cell before reading it to generate a random bit. This write operation sets a specific data state in the memory cell, and the subsequent read operation retrieves this state. The method uses the inherent physical variations in the memory cell's response to the write operation to generate randomness, while the preliminary write ensures that the cell is in a known state before measurement, thereby improving repeatability to above 95%.
2Adaptability or versatility
If physical variations in devices are utilized to create unique digital signatures, then each device has an unclonable identifier, but the digital signature lacks predictability and security due to bit changes
Solution Approach 1:
The patent employs feedback by comparing the random bit read from the first memory cell with a previously stored value or expected pattern. Based on this comparison, the system can detect and correct bit errors, ensuring that the digital signature maintains high predictability and security. The feedback mechanism allows the system to learn from physical variations and compensate for unpredictable bit changes, thereby resolving the contradiction between uniqueness and predictability.
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 approach significantly increases the repeatability of the random number generation to above 95%, making the digital signature highly repeatable and difficult to predict, thereby enhancing the security of the PUF system.
Implementation Method 1
The memory comprises a block of ferroelectric memory cells
Data Source
AI summary
A method of generating a random number from an electronic circuit memory and/or a system with the electronic circuit memory. The memory comprises a block of ferroelectric two transistor, two capacitor (2T-2C), memory cells. The method comprises: (i) first, writing a predetermined programming pattern to the block cells in a one transistor, one-capacitor (1T-1C) mode, thusly writing, per cell, a same data state to both a first and second sub-cell of the cell; (ii) second, reading the cells in a 2T-2C mode to generate a random number comprising a random bit from each of the cells; (iii) third, restoring the random number into the cells in a 2T-2C mode, thusly writing, per cell, a complementary data state to both a first and second sub-cell of the cell, responsive to a respective random number bit; and fourth, imprinting the random number in each cell in the block.