Semiconductor Memory Data Buffering and Error Propagation Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional semiconductor memory devices face challenges in preventing write errors from propagating to stored data, especially when data is not buffered, leading to data corruption and inability to recover data during power disconnection or errors.
Innovation Solution
A semiconductor memory system with a host device and memory manager that includes a data buffer, block boundary manager, and error propagation detector to manage data writing, detect errors, and reissue write commands to prevent data destruction, utilizing a logical-physical conversion table to handle multi-level flash memory blocks and requesting new blocks for secure writing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is written to nonvolatile memory without buffering, then writing speed is improved, but write errors can propagate to stored data causing data corruption
Solution Approach 1:
The patent segments the data writing process into distinct phases: pre-write buffering of data, error detection during writing, and selective re-writing based on error propagation detection. This segmentation allows the system to maintain high writing speeds while preventing error propagation to stored data by isolating potential errors to specific segments that can be corrected without affecting other data.
Solution Approach 2:
The patent introduces an intermediary error propagation detection mechanism between the data buffer and the nonvolatile memory. This intermediary detects whether write errors propagate to stored data and provides feedback to control the re-writing process, thereby preventing data corruption while maintaining efficient writing operations.
2Reliability
If data is buffered during writing, then error propagation can be prevented, but data recovery capability is reduced when power disconnection occurs
Solution Approach 1:
The patent implements preliminary buffering of data before writing to nonvolatile memory. This preliminary action ensures that data is prepared and protected in a recoverable state, enabling data recovery even if power disconnection occurs during the writing process. The buffer acts as a safety net that preserves data integrity.
Solution Approach 2:
The patent employs a feedback mechanism where the error propagation detector monitors the writing process and provides feedback to the control system. This feedback enables the system to detect errors, determine their impact on stored data, and initiate appropriate recovery procedures, thereby maintaining both error prevention and data recoverability.
3Quantity of substance
If multi-level flash memory is used to increase capacity, then storage density is improved, but error propagation to stored data becomes more complex
Solution Approach 1:
The patent applies local quality by detecting and addressing error propagation at specific local levels within the multi-level flash memory structure. Instead of treating all memory cells uniformly, the system identifies which specific cell levels are affected by write errors and applies targeted correction only to those local areas, thereby managing the complexity introduced by multi-level storage.
Solution Approach 2:
The patent introduces dynamic error propagation detection and correction mechanisms that adapt to the specific conditions of multi-level flash memory operations. The system dynamically determines the extent of error propagation based on the current writing operation and memory state, allowing flexible management of complexity in high-capacity storage systems.
Data Source
AI summary
A host device 200A includes a data buffer 250. When data has been already written to a part of a physical block and data is additionally written to the physical block, it is determined whether or not the data written to the physical block is held in the data buffer. When the data is held, data is written to the block, and when an error exists, data in unit of physical blocks is rewritten. When the data is not held in the data buffer, a new physical block is required to be secured and then, data is written to the new block. Thereby, even when power is shut off or an error occurs during writing in the semiconductor memory device, destruction of data already written is prevented.