Zoned Storage Garbage Collection Using Cache Pool Aggregation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Garbage collection in zoned storage devices (ZNS) affects performance due to increased query requests and input/output operations, especially as NAND capacity increases, leading to pressure on the device and reduced reading and writing efficiency.

Innovation Solution

A method and device for garbage collection in ZNS devices that determine valid data segments using file object metadata, aggregate them into cache pools, and write them sequentially into target zones, reducing metadata queries and IOPS.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional garbage collection is performed in SSD, then data can be cleaned up, but channel bandwidth is consumed significantly and performance is reduced

Engineering Contradiction:

Improvedata cleanup capabilityVSAvoidSSD performance

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent introduces a host-based garbage collection mechanism where the host acts as an intermediary to manage and clean up invalid data blocks in the SSD. Instead of the SSD performing garbage collection internally (which consumes channel bandwidth), the host identifies invalid blocks through metadata and issues trim commands to the SSD, thereby eliminating the performance penalty while maintaining data cleanup capability

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent inverts the traditional garbage collection approach by moving the garbage collection function from the SSD (device-side) to the host (controller-side). Traditionally, the SSD performs GC internally; this patent reverses that by having the host perform GC and communicate results to the SSD via trim commands, thus resolving the bandwidth consumption issue

Inventive Principle:

Principle #13The other way round (Inversion)

2Reliability

If garbage collection is performed inside the SSD, then storage space is reclaimed, but delay of the host is not ensured

Engineering Contradiction:

Improvestorage space reclamationVSAvoidhost delay

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The host serves as an intermediary that coordinates garbage collection operations, allowing the SSD to perform GC in the background without blocking host operations. The host issues GC commands and receives notifications, ensuring that critical host operations are not delayed while storage space is reclaimed through coordinated background processing

Inventive Principle:

Principle #24Intermediary (Mediator)

3Reliability

If garbage collection is performed frequently, then storage space is maintained, but lifespan of the SSD is affected

Engineering Contradiction:

Improvestorage space maintenanceVSAvoidSSD lifespan

Core Design Contradiction:

ReliabilityVSDuration of action of stationary object

Solution Approach 1:

The patent implements a feedback mechanism where the host monitors SSD health metrics and storage space utilization, dynamically adjusting the frequency and intensity of garbage collection operations. When the SSD shows signs of wear or when storage space is adequately maintained, GC frequency is reduced, thereby preserving SSD lifespan while still ensuring storage space maintenance through intelligent, adaptive control

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20260003785A1Method and device for garbage collection of a zoned storage device, and storage medium

Publication Date: 2026.01.01 BEIJING VOLCANO ENGINE TECH CO LTD

AI summary

Provided in examples of the disclosure are a method and device for garbage collection of a zoned storage device, and a storage medium. Valid data segments in a source zone to be collected in the zoned storage device are determined, the read valid data segments are written into cache pools, and the valid data segments in the same cache pool are aggregated into a valid aggregated data segment; and a target cache pool to be written into the zoned storage device is determined, the valid aggregated data segment in the target cache pool is written into a target zone, and storage space of the source zone is collected.