L2P Table Segmentation for SSD Power Loss Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In memory systems like solid-state drives (SSDs), restoring power after unexpected loss leads to a long time for rebuilding the logical-to-physical address translation table (L2P table), causing delays in read/write access and potential loss of write data due to limited power storage.
Innovation Solution
A memory system with a controller that divides the L2P table into table areas, periodically writes address translation information and update logs to nonvolatile memory blocks, and saves lists of lost logical addresses. Upon power restoration, the controller rebuilds the address translation information and updates it to reflect lost data, allowing for quicker access and data integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the entire L2P table is rebuilt after power loss, then data integrity is restored, but the time required for read/write access increases significantly
Solution Approach 1:
The L2P table is divided into multiple table areas, each corresponding to a specific LBA range. The controller selectively rebuilds only the affected table area instead of the entire table, significantly reducing the rebuild time while maintaining data integrity for the impacted portion.
Solution Approach 2:
The controller periodically saves the L2P table to nonvolatile memory before power loss occurs. This preliminary action ensures that when power is restored, the table can be quickly restored from the saved version without requiring a complete rebuild, thus reducing access time while maintaining integrity.
2Reliability
If the L2P table is saved to nonvolatile memory, then data loss is prevented, but the quantity of writable data is limited by capacitor storage
Solution Approach 1:
The L2P table is segmented into multiple table areas, each saved to a separate nonvolatile memory block. This segmentation allows the controller to save only the affected portions after power loss, reducing the total data quantity that needs to be written while maintaining comprehensive data protection.
Solution Approach 2:
The controller determines which specific table areas need to be saved based on the power loss conditions and only saves those specific portions to nonvolatile memory. This local approach optimizes the use of capacitor storage by writing only the necessary data portions rather than the entire table.
3Reliability
If the controller waits for complete L2P table rebuild before allowing access, then data consistency is ensured, but host waiting time increases
Solution Approach 1:
The L2P table is divided into multiple table areas that can be rebuilt independently. The controller can allow host access to unaffected table areas while continuing to rebuild only the affected areas in the background, ensuring data consistency for critical data while reducing overall host waiting time.
Solution Approach 2:
The controller dynamically adjusts the rebuild process by identifying which table areas are affected by power loss and prioritizing their rebuild. This dynamic approach allows the system to maintain data consistency for critical data while minimizing the time hosts need to wait for access.
Data Source
AI summary
According to one embodiment, in response to restoration of power to a memory system, a controller in the memory system notifies a host that the memory system is ready. When an input/output command specifying a logical address belonging to a logical address range is received, the controller selects a block corresponding to the logical address range and rebuilds, based on address translation information and an update log which are stored in the selected block, the latest address translation information corresponding to the logical address range. The controller updates the rebuilt latest address translation information, based on a list of logical addresses corresponding to lost write data, stored in the selected block.