Memory Sub-System Controller Data Compaction for Zoned Namespaces

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Memory devices face challenges in managing data compaction for zones, leading to issues such as slow charge loss and increased latency due to prolonged maintenance of open blocks.

Innovation Solution

The memory sub-system controller manages data compaction by opening SLC blocks for specific zones, tracking zone-to-block relationships, and performing media management operations to copy data from SLC blocks to QLC blocks when capacity thresholds are met, thereby reducing slow charge loss.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If SLC blocks are kept open for prolonged periods to maintain data, then data retention is improved, but slow charge loss increases and latency increases

Engineering Contradiction:
Improvedata retentionVSAvoidslow charge loss
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The system performs preliminary compaction of zone data into SLC blocks before they are fully filled, and proactively copies data to QLC blocks when capacity thresholds are approached. This preliminary action prevents the need to keep SLC blocks open indefinitely, thereby reducing slow charge loss while maintaining data retention.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system introduces QLC blocks as an intermediary storage medium between SLC blocks and permanent storage. Data is temporarily held in SLC blocks during active operations, then systematically transferred to QLC blocks which serve as a buffer, allowing SLC blocks to be closed and erased sooner without compromising data availability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If SLC blocks are kept open for prolonged periods to maintain data, then data retention is improved, but latency increases

Engineering Contradiction:
Improvedata retentionVSAvoidlatency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary compaction and data organization into SLC blocks, then proactively initiates copy operations to QLC blocks when capacity thresholds are approached. This preliminary action eliminates the need to maintain open blocks indefinitely, reducing the time blocks remain open and thereby reducing latency while preserving data retention.

Inventive Principle:
Principle #10Preliminary action

3Loss of energy

If data is compacted frequently to reduce slow charge loss, then slow charge loss is reduced, but CPU overhead increases

Engineering Contradiction:
Improveslow charge lossVSAvoidCPU overhead
Core Design Contradiction:
Loss of energyVSDevice complexity

Solution Approach 1:

Instead of performing full compaction operations continuously, the system applies partial compaction only when zone data reaches specific capacity thresholds. This selective, threshold-based approach reduces the frequency and intensity of compaction operations, thereby lowering CPU overhead while still effectively managing slow charge loss through targeted data movement to QLC blocks.

Inventive Principle:
Principle #16Partial or excessive action

4Loss of energy

If zone data is copied to QLC blocks proactively, then slow charge loss is reduced, but device complexity increases

Engineering Contradiction:
Improveslow charge lossVSAvoidmedia management
Core Design Contradiction:
Loss of energyVSDevice complexity

Solution Approach 1:

The system implements self-service media management where the memory device autonomously monitors SLC block capacity thresholds and automatically initiates data copy operations to QLC blocks without requiring external controller intervention. This self-service approach manages the increased device complexity internally, allowing the system to proactively reduce slow charge loss while maintaining a clean interface with external controllers.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS20250036319A1Managing data compaction for zones in memory devices
Publication Date: 2025.01.30 MICRON TECHNOLOGY INC
  • US20250036319A1 patent drawing
  • US20250036319A1 patent drawing
  • US20250036319A1 patent drawing

AI summary

A system and method for managing data compaction in zones in memory devices. An example method includes responsive to determining that a capacity counter associated with a first zoned namespace (ZNS) zone satisfies a threshold criterion, causing a memory device to copy, from a first management unit of the memory device to a second management unit of the memory device, first data associated with the first ZNS zone, wherein the first management unit comprises a first subset of memory pages configured to store a first number of bits per memory cell and the second management unit comprises a second subset of memory pages configured to store a second number of bits per memory cell; and causing the memory device to copy, from the first management unit to a third management unit of the memory device, second data associated with a second ZNS zone.