SSD Metadata Management via Contiguous File Node Grouping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The 'copy-on-write' approach in solid state storage devices (SSDs) leads to performance encumbrances when managing frequently-modified data with small file metadata, as it involves small and isolated write operations, burdening the SSD with inefficient transaction handling.

Innovation Solution

Implementing a method where a computing device analyzes file nodes against temporal transaction information to identify groups that were previously modified, modifies them, allocates contiguous storage space, and writes these groups contiguously, optimizing metadata management within SSDs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If copy-on-write approach is implemented in SSDs, then data integrity and reliability are improved, but performance deteriorates when managing frequently-modified small metadata files

Engineering Contradiction:
Improvedata integrityVSAvoidtransaction handling performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent merges multiple small metadata write operations into a single contiguous write operation by identifying temporally-related file nodes and allocating them to the same SSD block. This combining approach maintains the copy-on-write reliability while eliminating the performance penalty of multiple small writes by treating them as one unified operation.

Inventive Principle:
Principle #5Merging (Combining)

2Device complexity

If small metadata files are stored separately in SSD blocks, then data organization simplicity is improved, but write efficiency deteriorates due to multiple isolated operations

Engineering Contradiction:
Improvedata organizationVSAvoidwrite efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent introduces a new dimension to data organization by adding temporal relationship tracking alongside traditional spatial storage. File nodes are organized not only by their storage location but also by their temporal modification patterns, allowing the system to group and write related metadata together without complicating the underlying storage structure.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Productivity

If contiguous storage space is allocated for metadata groups, then write operation efficiency is improved, but storage management complexity increases

Engineering Contradiction:
Improvewrite operation efficiencyVSAvoidstorage management
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by pre-identifying groups of temporally-related file nodes before the actual write operation occurs. By analyzing temporal transaction information in advance and pre-allocating contiguous storage space for identified groups, the system prepares the storage structure beforehand, which simplifies the execution phase and improves write efficiency without proportionally increasing management complexity.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10740015B2Optimized management of file system metadata within solid state storage devices (SSDs)
Publication Date: 2020.08.11 APPLE INC
  • US10740015B2 patent drawing
  • US10740015B2 patent drawing
  • US10740015B2 patent drawing

AI summary

The embodiments set forth a technique for managing file nodes within a storage device. According to some embodiments, the method can include the steps of (1) receiving a transaction request that involves modifying a plurality of file nodes, (2) analyzing the plurality of file nodes against temporal transaction information to identify, among the plurality of file nodes, at least one group of file nodes that were previously modified in conjunction with at least one previously-executed transaction request, (3) modifying each file node in the at least one group of file nodes in accordance with the transaction request to establish a modified group of file nodes, (4) allocating, within the storage device, physical storage space for storing the modified group of file nodes, and (5) contiguously writing the modified group of file nodes into the allocated physical storage space.