NVDIMM Block Reclamation for Wear Leveling and Memory Availability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Non-volatile dual in-line memory modules (NVDIMMs) face issues with poor fatigue resistance and limited write times, leading to unbalanced wear among memory blocks, which affects their service life and performance.
Innovation Solution
A memory block reclamation method and apparatus that involves a medium controller receiving an unmap command with memory logical address information to reclaim memory blocks, deleting the mapping relationship, and adding the physical address to an idle list, enhancing wear leveling and improving storage performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If memory blocks are not reclaimed promptly, then wear leveling is improved, but memory availability and reclamation speed deteriorate
Solution Approach 1:
The medium controller proactively identifies and reclaims memory blocks before they reach maximum write times. By monitoring write counts and initiating reclamation operations in advance, the system prevents wear imbalance while maintaining high memory availability. The controller sends unmap commands to reclaim memory blocks that are no longer needed, converting them to available blocks before they become problematic.
Solution Approach 2:
The system implements continuous monitoring of memory block write times and wear levels. Based on this feedback, the medium controller dynamically adjusts reclamation operations, identifying blocks that need reclamation and executing unmap commands accordingly. This closed-loop control ensures optimal wear leveling while maintaining memory performance.
2Duration of action of stationary object
If memory management operations are increased, then service life is prolonged, but system complexity increases
Solution Approach 1:
The medium controller autonomously manages memory block reclamation without requiring complex host intervention. It independently monitors write times, identifies blocks for reclamation, and executes unmap commands. This self-managing approach extends NVDIMM service life while keeping the control logic centralized and manageable within the medium controller.
Solution Approach 2:
The patent combines wear leveling, garbage collection, and memory management functions into a unified medium controller. By merging these operations under a single control entity that communicates via NVDIMM-P protocol, the system achieves comprehensive memory management without the complexity of multiple independent management systems.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A memory block reclamation method and apparatus are provided. The method includes: receiving, by a medium controller, an unmap command sent by a memory controller, where the unmap command includes memory logical address information of a to-be-reclaimed memory block (S201); and reclaiming, by the medium controller, the memory block based on the memory logical address information, where the reclaimed memory block is an available memory block (S202). After receiving the unmap command sent by the memory controller, the medium controller releases a corresponding memory block and reclaims the memory block, based on the memory logical address information of the to-be-reclaimed memory block included in the unmap command, thereby enhancing a medium management operation effect of the medium controller, improving storage performance of an NVDIMM, and prolonging a service life of the NVDIMM.