Storage Device Retry Reads for V-NAND Retention Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing read operations in V-NAND flash memory devices suffer from increased raw bit error rate (RBER) and multiple retry reads due to data retention, leading to degraded read performance and Quality of Service (QoS) latency.
Innovation Solution
Performing data retry read operations in the hardware layer of the storage device, rather than the application layer, and prioritizing them when uncorrectable errors are detected, thereby reducing latency and improving read performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple retry read operations are performed to correct retention errors, then data reliability is improved, but read performance and QoS latency are worsened
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing multiple candidate read offsets (first read offset, second read offset, third read offset) in the storage device before actual read operations occur. When retention errors are detected, the system can immediately switch between these pre-prepared offsets without performing sequential trial reads, thereby reducing latency while maintaining data reliability through multiple correction attempts.
2Reliability
If retry read operations are performed at the application layer, then error correction capability is improved, but device complexity and processing overhead are worsened
Solution Approach 1:
The patent extracts the retry read operation logic from the application layer and implements it directly within the storage device's hardware layer. The storage device autonomously manages multiple read attempts using pre-stored candidate offsets, eliminating the need for complex application-layer coordination and reducing overall system processing overhead while maintaining robust error correction capabilities.
3Reliability
If multiple layer interactions are used for read operations, then data reliability is improved, but read latency is worsened
Solution Approach 1:
The patent merges the reliability functions across multiple layers into a unified approach where the storage device's hardware layer directly performs multiple read operations with different offsets. This consolidation eliminates the need for sequential layer-by-layer interactions and coordination overhead, achieving both high data reliability through multiple correction attempts and low latency by executing operations directly at the hardware level.
Data Source
AI summary
A method for performing data read operation on a storage device, includes: based on receiving a command from a host device, generating a read request using a first layer associated with the storage device, wherein the read request includes at least one memory physical parameter, a default read offset, and at least one retry read offset; communicating, via a memory interface layer, the read request to a second layer associated with the storage device; performing a default read operation in the second layer based on the at least one memory physical parameter and the default read offset; detecting a first error status of the read request during the default read operation in the second layer; and based on the first error status of the read request indicating the first uncorrectable error, performing a retry read operation in the second layer using a retry read offset.