Replication Endpoint Host Mapping for Redundant Write Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional storage systems face inefficiencies in data management and reliability due to redundant write operations and lack of centralized control over flash drives, leading to increased latency and unreliability issues.

Innovation Solution

The implementation of direct data storage systems using direct-mapped flash storage systems addresses data blocks within the flash drives, utilizing a centralized storage system that addresses data blocks within the flash drives directly and without address translation by the storage controllers, thereby reducing redundant write operations and enhancing reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If storage controllers perform address translation for flash drives, then data management flexibility is improved, but write operations become redundant and reliability decreases

Engineering Contradiction:
Improvedata management flexibilityVSAvoidflash drive reliability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent extracts the address translation function from the storage controller and implements it directly in the flash drive controller. This removes the redundant translation layer in the data path, allowing flash drives to be addressed directly without controller-mediated translation, thereby improving reliability while maintaining management flexibility through the centralized metadata server.

Inventive Principle:
Principle #2Taking out (Extraction)

2Ease of operation

If storage controllers mediate all data access to flash drives, then data management control is improved, but access latency increases

Engineering Contradiction:
Improvedata management controlVSAvoiddata access latency
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The patent introduces a centralized metadata server as an intermediary that manages data location information and flash drive mapping. This allows storage controllers to maintain operational control through metadata management while enabling direct data paths between hosts and flash drives, significantly reducing access latency by eliminating controller-mediated address translation for every data operation.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Device complexity

If redundant write operations are performed through storage controllers, then data management overhead is reduced, but system performance and productivity decrease

Engineering Contradiction:
Improvedata management overheadVSAvoidsystem performance
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent segments data management functions into two distinct components: metadata management (handled by the centralized metadata server) and data operations (handled directly by flash drive controllers). This segmentation eliminates redundant write operations through storage controllers while maintaining comprehensive data management control, thereby improving system performance and productivity without increasing overall management complexity.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS12517873B2Mapping equivalent hosts at distinct replication endpoints
Publication Date: 2026.01.06 PURE STORAGE INC
  • US12517873B2 patent drawing
  • US12517873B2 patent drawing
  • US12517873B2 patent drawing

AI summary

A method of mapping equivalent hosts at distinct replication endpoints, the method including: receiving an indication that a first host mapped to a source dataset at a replication source is equivalent to a second host for a replica dataset at a replication target of the source dataset; transmitting, from the replication source to the replication target and based on an update to the source dataset received from the first host, metadata describing the update to the source dataset; and mapping, based on the indication that the first host mapped to the source dataset at the replication source is equivalent to the second host for the replica dataset at the replication target, the second host to the replica dataset at the replication target of the source dataset.