Mapping Table Update via Partial Garbage Collection Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for updating logical-to-physical address mapping tables in data storage devices during garbage collection procedures lead to performance degradation due to the need for frequent access to non-volatile memory, especially when the data buffer is small, causing delays in processing write/read commands from hosts.
Innovation Solution
Implementing a partial garbage collection procedure that segments the copying of valid data from source blocks to destination blocks, allowing for real-time updating of the mapping table based on the logical and physical addresses of copied data, and maintaining a backup table to manage the mapping relationships efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the mapping table is updated after the destination block is fully filled during garbage collection, then the mapping table update is simple and batched, but the data storage device cannot process host commands during garbage collection, causing performance degradation and delays
Solution Approach 1:
The patent segments the mapping table update process into multiple incremental updates performed during different stages of the garbage collection procedure. Instead of waiting for the destination block to be fully filled before updating, the mapping table is updated in parts as data is copied from source blocks to the destination block, allowing the system to maintain command processing capability while performing garbage collection.
Solution Approach 2:
The patent performs preliminary actions by pre-positioning data in a data buffer during garbage collection operations. The mapping table is updated with preliminary information about data locations, and the actual complete update is finalized after the destination block is filled. This allows the system to accept and buffer host commands during garbage collection without losing productivity.
2Productivity
If the mapping table is updated frequently during partial garbage collection, then command processing is maintained, but the frequent access to non-volatile memory increases overhead and complexity
Solution Approach 1:
The patent introduces a data buffer as an intermediary layer between the host and the non-volatile memory. The buffer temporarily stores data and mapping information during garbage collection operations, reducing the need for frequent direct accesses to the non-volatile memory. This intermediary mechanism simplifies the overall system complexity while maintaining high command processing capability.
Solution Approach 2:
The patent uses copying mechanisms where the mapping table information is copied to the data buffer during garbage collection, and then the buffer is used to manage subsequent updates. This copying approach reduces the complexity of direct mapping table management by delegating update operations to the buffer, which can handle changes more efficiently before committing to the non-volatile memory.
3Reliability
If the destination block is fully filled before updating the mapping table, then the update operation is atomic and simple, but the system cannot respond to host commands during the filling process, reducing overall throughput
Solution Approach 1:
The patent implements feedback mechanisms where the mapping table is updated with partial information during the garbage collection process, and subsequent updates are adjusted based on the actual data placement. The system continuously monitors the garbage collection progress and adjusts mapping table updates accordingly, ensuring accuracy while maintaining high throughput by not waiting for complete block filling.
Solution Approach 2:
The patent makes the mapping table update process dynamic by allowing it to adapt to the progress of garbage collection. Instead of a static wait-for-completion approach, the system dynamically updates the mapping table at appropriate intervals during the garbage collection process, balancing between update accuracy and system responsiveness to host commands.
Data Source
AI summary
A mapping table updating method executable by a data storage device is provided. The data storage device includes a non-volatile memory and a controller. The mapping table updating method includes steps of: step A: configuring the controller to process a command issued by a host, and determine whether to trigger a partial garbage collection procedure when the command is a write command; when it is determined to trigger the partial garbage collection procedure, then performing step B: copying partial valid data in at least one source block to a destination block according to a segmentation condition; and step C: updating a logical-to-physical address mapping table of the data storage device according to a logical address of the copied partial valid data and a physical address in the destination block where the partial valid data is located, and returning to perform the step A.