Storage Drive Peer Grouping for Slow Response Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In large data storage systems, a slow hard drive can significantly impact overall performance if not diagnosed and addressed promptly, as operations are shared among multiple drives, and existing monitoring systems may fail to detect or diagnose such issues effectively.

Innovation Solution

A method and system for detecting slow storage drives by dynamically creating peer groups of drives, comparing performance statistics, passively monitoring response times, actively measuring workload metrics, and performing remedial actions such as repair, replacement, or removal of the slow drive based on threshold comparisons.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a slow hard drive is not diagnosed or diagnosed incorrectly and operations continue, then the storage system can maintain continuous operation, but the overall performance of the storage system is slowed down

Engineering Contradiction:
Improvestorage system performanceVSAvoiddrive performance consistency
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent segments the storage drive population into peer groups based on performance characteristics, allowing individual drives to be identified and addressed without impacting the entire system. This enables targeted remediation of slow drives while maintaining overall system operation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system implements continuous monitoring of drive performance metrics and uses this feedback to identify outlier drives. The feedback loop enables dynamic detection and classification of slow drives, allowing the system to adapt to changing performance conditions and maintain optimal operation.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If monitoring systems are used to detect drive degradation, then drive failures can be identified, but existing monitoring systems may fail to detect or diagnose slow drives effectively

Engineering Contradiction:
Improvedrive performance detection accuracyVSAvoidslow drive diagnosis complexity
Core Design Contradiction:
Measurement precisionVSDifficulty of detecting and measuring

Solution Approach 1:

The patent implements dynamic peer group formation that adapts to changing drive performance characteristics. The system continuously re-evaluates and re-groups drives based on current performance metrics, enabling accurate detection of slow drives even as system conditions change over time.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system monitors multiple performance parameters including response time, throughput, and error rates to identify slow drives. By analyzing changes in these parameters over time and comparing against peer groups, the system achieves precise detection and diagnosis of drive performance issues.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If one slow drive impacts the overall performance of the storage system, then system-wide performance degradation occurs, but identifying and addressing the specific slow drive is challenging

Engineering Contradiction:
Improveoverall storage system performanceVSAvoidslow drive identification ease
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent introduces peer groups as an intermediary layer between individual drives and system-wide performance. By comparing drives within their peer groups and identifying outliers, the system can pinpoint specific slow drives without requiring complex system-wide analysis, making identification and remediation straightforward.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS10481828B2Slow drive detection
Publication Date: 2019.11.19 SEAGATE TECH LLC
  • US10481828B2 patent drawing
  • US10481828B2 patent drawing
  • US10481828B2 patent drawing

AI summary

Implementations described and claimed herein provide a method and system for detecting slow storage drives. In one implementation, the method includes dynamically creating a storage drive peer group including a plurality of storage drives, comparing performance statistics of the storage drives in the storage drive peer group, selecting an outlier storage drive of the storage drive peer group based on the comparison of the performance statistics, passively monitoring response times of the storage drives in the storage drive peer group, comparing average response times of the storage drives in the storage drive peer group, flagging an outlier storage drive of the storage drive peer group with an outlier storage drive designation responsive to comparison of the average response times, actively measuring workload metrics of the outlier storage drive, comparing workload metrics data of the outlier storage drive to workload metrics reference data, and performing a remedial action.