Nonvolatile Memory Seed Selection for Data Randomization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current nonvolatile memory systems face challenges in maintaining data randomness, leading to potential errors and reliability issues due to lack of adequate randomness in programming data, particularly in adjacent memory cells.
Innovation Solution
A nonvolatile memory system with a vertical NAND structure that includes a data generating unit for creating random reference values, a seed selecting unit for choosing seeds from pre-stored tables or polynomials, and a randomizer to generate and program randomized data, ensuring increased randomness by regenerating reference values and selecting different seeds for each program request.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is programmed into memory cells without sufficient randomization, then the programming process is simple and fast, but data randomness is insufficient leading to errors and reliability issues
Solution Approach 1:
The patent pre-generates and stores multiple seeds in seed tables before the actual data programming operation. During programming, the system selects from pre-prepared seeds rather than generating them in real-time, which simplifies the overall process while ensuring adequate randomization for reliable data storage.
Solution Approach 2:
The patent introduces a seed selection mechanism as an intermediary between the data programming operation and the actual data writing. This seed selection unit chooses appropriate seeds from stored tables and passes them to the randomizer, which then generates the final randomized data. This intermediary layer ensures reliability without requiring complex real-time random number generation.
2Reliability
If the same seed is used for multiple program requests, then the system is simple to operate, but data randomness decreases leading to deep erase phenomena and errors
Solution Approach 1:
The patent implements a dynamic seed selection mechanism where the seed chosen for each program request depends on the current program counter value and the data being programmed. This dynamic approach ensures that different seeds are selected for different programming operations, maintaining data randomness and preventing deep erase phenomena while adapting to varying operational conditions.
Solution Approach 2:
The system uses feedback from the program counter and current data characteristics to determine which seed to select from the pre-stored tables. This feedback mechanism ensures that the seed selection is adaptive and context-dependent, maintaining randomness without requiring complex manual intervention or operation.
3Reliability
If reference values are regenerated for each program request, then data randomness increases preventing deep erase phenomena, but processing time increases
Solution Approach 1:
The patent pre-generates multiple reference values and stores them in seed tables before actual programming operations occur. During programming, the system retrieves appropriate reference values from these pre-prepared tables based on the program counter and data characteristics, rather than generating them in real-time. This preliminary action significantly reduces processing time while maintaining adequate data randomness to prevent deep erase phenomena.
Data Source
AI summary
A nonvolatile memory device includes a data generating unit for generating a first reference value randomly or pseudo-randomly according to a first program request to program data in a memory cell, a seed selecting unit for selecting at least one of a plurality of seeds using the first reference value, and a randomizer for generating randomized data by using the selected seed. The data generating unit regenerates the first reference value as a second reference value different from the first reference value when a second program request is made, and the seed selecting unit selects another seed using the second reference value.