Integrated Log Analytics for Flash Write Anomaly Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional storage systems face inefficiencies in managing flash storage due to redundant write operations and lack of direct control by the operating system, leading to reduced reliability and increased wear on flash drives.

Innovation Solution

A direct-mapped flash storage system where the operating system directly addresses data blocks without translation by storage controllers, managing allocation units and performing operations like data rewriting and erasure, thereby reducing redundant writes and enhancing reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If storage controllers translate and manage flash addresses, then data management is simplified, but redundant write operations increase and reliability decreases

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

Solution Approach 1:

The patent extracts the address translation function from the storage controller and places it directly in the operating system. The OS maintains its own mapping tables and directly translates logical addresses to physical flash addresses, eliminating the controller's address translation layer that caused redundant writes and reliability issues.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The operating system serves itself by implementing its own flash address translation and management mechanisms. The OS directly manages allocation units, performs garbage collection, and handles wear leveling without relying on storage controller services, thereby reducing redundant operations and improving reliability.

Inventive Principle:
Principle #25Self-service

2Ease of operation

If storage controllers manage allocation and rewriting, then operation is easier, but flash drive wear increases and lifespan decreases

Engineering Contradiction:
Improveflash management simplicityVSAvoidflash drive lifespan
Core Design Contradiction:
Ease of operationVSDuration of action of stationary object

Solution Approach 1:

The patent removes the allocation and rewriting management functions from the storage controller and transfers them to the operating system. The OS directly manages allocation units, tracks valid data locations, and performs rewriting operations, eliminating the controller's management layer that contributed to excessive wear.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The operating system performs self-service flash management by implementing its own allocation unit tracking, valid data identification, and rewriting mechanisms. This direct management approach optimizes write operations and reduces unnecessary flash modifications, extending drive lifespan.

Inventive Principle:
Principle #25Self-service

3Adaptability or versatility

If multiple layers of address translation are used, then data management is more flexible, but write operations become redundant and efficiency decreases

Engineering Contradiction:
Improvedata management flexibilityVSAvoidwrite operation efficiency
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent extracts the address translation function from the storage controller layer and consolidates it into a single translation layer within the operating system. This eliminates the multiple-layer translation architecture that caused redundant writes, while the OS maintains flexibility through its own mapping tables and management mechanisms.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS20260111420A1Flash system with integrated log analytics and anomaly detection
Publication Date: 2026.04.23 PURE STORAGE INC
  • US20260111420A1 patent drawing
  • US20260111420A1 patent drawing
  • US20260111420A1 patent drawing

AI summary

Log data generated by a computing system is received. An analytics operation is executed on the log data. Pattern recognition associated with the log data is applied to identify anomalous behaviors of the computing system.