Memory Data Shredding via Random Value Overprogramming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for shredding data stored in memory units, such as overprogramming with all zeros, fail to guarantee complete erasure of sensitive data, leaving it vulnerable to retrieval by unauthorized parties.
Innovation Solution
The method involves generating a random value and its complement, and then overprogramming the memory unit to write these values successively, effectively shredding the original data by flipping all bits irreversibly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is shredded by overprogramming with all zeros, then the process is simple and fast, but the data erasure is incomplete and vulnerable to retrieval
Solution Approach 1:
The data erasure process is segmented into multiple sequential programming operations with different patterns (e.g., all zeros, all ones, random patterns). Each programming pass targets different aspects of data remanence, ensuring comprehensive erasure through divided action rather than relying on a single programming operation
Solution Approach 2:
The patent implements periodic programming cycles with varying data patterns repeated across multiple passes. Each cycle programs the memory cell with a specific pattern, then subsequent cycles use different patterns to ensure complete bit flipping and eliminate any residual data traces that single-pass methods might miss
2Reliability
If data is shredded by multiple programming passes with different patterns, then data erasure completeness is improved, but the process time and complexity increase
Solution Approach 1:
The system performs preliminary assessment of the memory cell states before the full multi-pass shredding process. By initially programming with all zeros and then selectively applying additional patterns only where needed, the system reduces unnecessary programming operations while ensuring complete erasure, thereby minimizing overall time loss
Solution Approach 2:
The patent dynamically changes programming parameters such as pulse width, voltage amplitude, and pattern selection based on detected memory cell states. By adapting programming conditions in real-time, the system achieves complete data erasure with fewer passes compared to fixed-parameter methods, reducing the time required for the shredding process
Data Source
AI summary
Various examples in accordance with the present disclosure provide example methods, systems, and apparatuses that shred data stored in a memory unit.