Memory Controller Buffering for L2P Update Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In memory devices, especially those with limited working memory, the process of data removal marking through UNMAP commands can lead to significant latency due to the need to repeatedly load and update segmented logical-to-physical (L2P) data structures, impacting performance and longevity, particularly in scenarios with large or frequent data removal operations.
Innovation Solution
Implementing a buffer-based approach for small data removal commands and immediate processing for large commands, where small commands accumulate in a journal until full and then update the L2P data structure segments minimally, while large commands are processed directly without buffering to reduce segment loading and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data removal marking is performed by repeatedly loading and updating segmented L2P data structures, then data integrity is maintained, but significant latency occurs due to repeated segment loading
Solution Approach 1:
The patent applies preliminary action by pre-loading the entire L2P data structure into a buffer memory before data removal operations begin. This allows subsequent UNMAP commands to be processed by simply updating buffer entries rather than repeatedly loading segmented L2P data from storage, significantly reducing latency while maintaining data integrity through coordinated buffer updates and persistence mechanisms
Solution Approach 2:
The patent introduces a buffer memory as an intermediary between the segmented L2P data structure in storage and the data removal marking operations. The buffer acts as a working copy that can be rapidly updated without accessing slow storage media, serving as a mediator that decouples the speed requirements of UNMAP processing from the physical constraints of segmented L2P storage
2Productivity
If small data removal commands are buffered until full, then update performance improves by reducing segment loads, but command processing delay increases
Solution Approach 1:
The patent applies partial action by implementing a threshold-based buffering strategy where small UNMAP commands are accumulated in the buffer until a size threshold is reached, at which point a targeted update of only the affected L2P segments is performed. This partial updating approach improves productivity by reducing unnecessary segment loads while the threshold mechanism prevents excessive delay by triggering updates at appropriate intervals
3Loss of time
If large data removal commands are processed immediately without buffering, then command processing delay is reduced, but productivity decreases due to increased segment loading
Solution Approach 1:
The patent applies dynamics by making the data removal processing strategy adaptive based on command size. Small commands are routed to the buffer for batch processing, while large commands trigger immediate L2P segment loading and updating. This dynamic routing optimizes both productivity and response time by matching the processing approach to the characteristics of each UNMAP command
Data Source
AI summary
Devices and techniques for data removal marking in a memory device are described herein. A delete command can be received at the memory device. A count of data portions in the delete command can be compared to determine whether the count is below a threshold. In response to determining that the count of data portions is below the threshold, the data portions can be written to a buffer. When a buffer full event is detected, a segment of an L2P data structure can be loaded into working memory of the memory device. Then, each record in the segment of the L2P data structure that has a corresponding entry in the buffer can be updated to mark the data as removable (e.g., invalid).