Structured Dataset Conversion With Direct-Mapped Flash Storage

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing storage systems face inefficiencies in data management and reliability due to redundant write operations and lack of direct control over flash drives, leading to increased wear and reduced reliability.

Innovation Solution

Implementing a direct-mapped flash storage system where the operating system directly addresses data blocks without translation by storage controllers, offloading device management responsibilities, and utilizing non-volatile RAM for quick data buffering and power failure protection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If storage controllers translate and manage flash drive operations, then data management is simplified for the operating system, but device complexity increases and reliability decreases due to redundant write operations

Engineering Contradiction:
Improvedata management simplicityVSAvoidflash drive reliability
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent extracts the translation and management functions from the storage controller to the operating system. The OS directly addresses flash drive data blocks without controller intervention, eliminating the intermediate translation layer that causes redundant writes and reliability issues.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The operating system takes self-service responsibility for managing flash drive operations directly. The OS performs data block addressing and management without relying on storage controller translation, making the system self-sufficient and reducing points of failure.

Inventive Principle:
Principle #25Self-service

2Device complexity

If storage controllers manage flash drive operations, then device control is centralized, but wear increases due to redundant write operations

Engineering Contradiction:
Improvecontrol structureVSAvoidflash drive lifespan
Core Design Contradiction:
Device complexityVSDuration of action of moving object

Solution Approach 1:

The patent removes the storage controller's translation function and assigns it to the operating system. This extraction eliminates the redundant write operations that occur when controllers translate addresses, thereby reducing wear and extending flash drive lifespan.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent eliminates the storage controller as an intermediary between the operating system and flash drive. By establishing direct communication, the system removes the source of redundant writes that accelerate wear.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Ease of operation

If traditional storage systems are used without direct OS control, then device management is abstracted, but data integrity is compromised during power failures

Engineering Contradiction:
Improvedevice management abstractionVSAvoiddata integrity
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The operating system directly manages flash drive operations including power failure protection. This self-service approach allows the OS to implement its own data integrity mechanisms without relying on controller abstraction, ensuring data integrity during power failures.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent removes the storage controller intermediary that prevents direct OS control over data blocks. This direct control enables the OS to ensure data integrity during power failures through direct management of data writing and buffering.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS20260111154A1Converting unstructured datasets into structured datasets
Publication Date: 2026.04.23 PURE STORAGE INC
  • US20260111154A1 patent drawing
  • US20260111154A1 patent drawing
  • US20260111154A1 patent drawing

AI summary

Executing a machine learning model in an artificial intelligence infrastructure that includes one or more storage systems and one or more graphical processing unit (‘GPU’) servers, including: receiving, by a graphical processing unit (‘GPU’) server, a dataset transformed by a storage system that is external to the GPU server; and executing, by the GPU server, one or more machine learning algorithms using the transformed dataset as input.