Memory Controller Trim Handling via Data Pattern Flagging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory systems face challenges in efficiently managing and releasing unnecessary data areas, particularly when the area requested to be trimmed is smaller than the management unit, as they cannot set such areas to 'unallocated' without additional storage and special processing.
Innovation Solution
The memory system employs a 'specified data pattern' as a flag to indicate unnecessary areas, allowing for the release of management units by writing this pattern to areas not aligned to the management unit, thereby reducing unnecessary data movement and increasing usable storage space without requiring additional storage for flags.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the memory system uses traditional flag-based tracking for unallocated areas, then it can precisely manage delete areas smaller than management units, but it requires additional storage space and increases device complexity
Solution Approach 1:
The patent uses a specified data pattern (copy of a known pattern) written to the entire management unit as a substitute for traditional flag-based tracking. When a delete area smaller than a management unit occurs, the system writes a predetermined data pattern to the entire management unit containing that area, making it identifiable as unallocated without requiring separate flag storage.
Solution Approach 2:
The patent changes the state parameter of the management unit by writing a specific data pattern to it. This pattern-writing operation transforms the management unit's state to indicate it contains unallocated areas, allowing the system to track delete areas by content rather than by separate metadata flags.
2Device complexity
If the memory system writes specified data pattern to entire management units to track unallocated areas, then it eliminates need for additional flag storage, but it increases data movement overhead
Solution Approach 1:
The system uses the data pattern itself as the tracking mechanism, making the storage medium self-describing. The specified data pattern written to management units serves dual purposes: both as data content and as the indicator of unallocated status, eliminating the need for separate tracking structures and reducing overall system complexity.
Solution Approach 2:
The patent merges the data storage function with the metadata tracking function. The specified data pattern serves both as actual data content and as the indicator (replacing separate flags), combining two functions into a single mechanism to reduce device complexity.
3Productivity
If the memory system releases management units containing specified data pattern, then it increases usable storage space efficiency, but it requires additional processing to identify and release these units
Solution Approach 1:
The system implements a feedback mechanism where the specified data pattern written to management units creates a detectable state that triggers automatic identification and release processes. The controller periodically scans for management units containing the specified pattern and automatically releases them, creating a closed-loop system that improves storage efficiency without manual intervention.
Data Source
AI summary
According to one embodiment, a memory system includes nonvolatile memory and storage unit storing a translation table indicating, by a predetermined management unit, relationships between logical addresses specified by a host and physical addresses in the nonvolatile memory. A memory system of the embodiment includes a controller that when receiving from the host a delete notification indicating a delete area smaller than the management unit specified by a logical address, write a specified data pattern to an area of the nonvolatile memory having a physical address corresponding to the delete area.