UFS Write Booster Buffer Flush Resume After Partial Flush Failure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing flash storage devices face issues with partial flush errors during the transfer of data from a write booster buffer to normal storage, leading to synchronization problems between memory tables and increased error rates, which result in time-consuming re-synchronization and potential data loss.
Innovation Solution
A flash storage device system that includes a write booster buffer with flush progress tracking, allowing the host controller to be notified of flush termination, transmitting context information about successfully flushed data structures, and resuming the flush from a designated completion position.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the flush operation is interrupted due to an error or fault, then the system must stop and re-synchronize, but this results in time loss and potential data loss
Solution Approach 1:
The patent applies preliminary action by maintaining a progress indicator that continuously tracks the flush operation status before any error occurs. This progress indicator records which data structures have been successfully flushed, allowing the system to resume from the last known good state rather than restarting from scratch, thus preventing time loss while maintaining data integrity
Solution Approach 2:
The patent implements feedback through the progress indicator that provides continuous information about the flush operation's state. When an error interrupts the flush, the progress indicator feedback allows the system to identify exactly where to resume, enabling efficient recovery without full re-synchronization and reducing both time loss and data loss risks
2Productivity
If the system performs frequent buffer flushes to maintain performance, then operation performance improves, but the risk of encountering errors during flush increases
Solution Approach 1:
The progress indicator is maintained as a preliminary protective measure during flush operations, recording the state before errors can occur. This allows the system to safely perform frequent flushes for performance while having a built-in recovery mechanism that prevents catastrophic failure, thus maintaining both productivity and reliability
Solution Approach 2:
The progress indicator acts as a cushioning mechanism that prepares the system for potential errors during frequent flush operations. By having the recovery information ready in advance, the system can tolerate more frequent flush attempts without compromising reliability, thus enabling sustained high performance
3Reliability
If the system restarts the flush operation from the beginning after an error, then data consistency is ensured, but time consumption increases significantly
Solution Approach 1:
The progress indicator is maintained throughout the flush operation to record the exact point where data consistency is last achieved. After an error, this preliminary tracking information allows the system to resume from the consistent state rather than restarting, maintaining data consistency while dramatically reducing the time required for recovery
Solution Approach 2:
The progress indicator provides feedback about the flush state, enabling the system to identify the last consistent state after an error. This feedback mechanism allows selective re-execution of only the necessary portion of the flush operation, ensuring data consistency without the time penalty of a complete restart
Data Source
AI summary
Methods that may be performed by a universal flash storage (UFS) system of a computing device that provide a write booster buffer partial flush process and a flush resume/retry process in a flash storage device. The method for managing operations of a write booster buffer of a flash memory device includes notifying a host controller that a flush of the write booster buffer has been terminated, transmitting, to the host controller, context information describing one or more data structures that were successfully flushed. The method may include receiving a partial unmap command from the host controller to unmap the one or more data structures that were successfully flushed and may include resuming the flush of the write booster buffer.