Storage System Latency Correlation Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Identifying the significant contributing factors to storage system latency is a time-intensive and computationally expensive process, making it challenging to effectively address and optimize storage system performance.

Innovation Solution

Instrumenting storage systems with sensors to measure latency and component activity, correlating these measurements with storage system latency to identify potential contributing factors, and using correlation analysis to validate hardware upgrades or software patches.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional methods are used to identify contributing factors to storage system latency, then comprehensive analysis can be performed, but the process becomes time-intensive and computationally expensive

Engineering Contradiction:
Improvelatency factor identification accuracyVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts and isolates specific sensor measurements from the storage system components that are most likely to contribute to latency. By selecting and monitoring only the most relevant metrics (such as disk I/O, network activity, CPU utilization) rather than analyzing all possible system parameters, the solution achieves comprehensive latency factor identification while significantly reducing analysis time and computational resources required.

Inventive Principle:
Principle #2Taking out (Extraction)

2Measurement precision

If comprehensive sensor measurements are collected from all storage system components, then complete latency analysis is possible, but system complexity and measurement overhead increase

Engineering Contradiction:
Improvelatency measurement completenessVSAvoidsensor measurement system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies local quality by instrumenting only the specific storage system components and processes that are most likely to contribute to latency with sensors. Rather than uniformly monitoring all system elements, the solution selectively places measurement instruments on critical components (such as disk drives, network interfaces, and specific software processes), thereby achieving complete latency analysis capability while minimizing the overall complexity of the measurement system.

Inventive Principle:
Principle #3Local quality

3Measurement precision

If correlation analysis is performed on all sensor measurements to identify latency factors, then accurate bottleneck identification is achieved, but computational resources and processing time increase

Engineering Contradiction:
Improvebottleneck identification accuracyVSAvoidcomputational energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent implements partial action by performing correlation analysis only on the selected subset of sensor measurements that are most relevant to latency, rather than analyzing all available system metrics. By focusing computational resources on the most promising measurements (such as I/O operations, network throughput, and CPU utilization during latency events), the solution achieves accurate bottleneck identification while significantly reducing the computational energy consumption required for the analysis.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS10503408B2Deployment of an upgrade to a storage system based on correlation analysis of measurements of the storage system
Publication Date: 2019.12.10 HEWLETT PACKARD ENTERPRISE DEV LP
  • US10503408B2 patent drawing
  • US10503408B2 patent drawing
  • US10503408B2 patent drawing

AI summary

Described herein are methods, systems and machine-readable media that facilitate an analysis of the contributing factors of storage system latency. The variation over time of the storage system latency is measured, along with the variation over time of the activity of various processes and/or components, the various processes and/or components being potentially contributing factors to the storage system latency. The latency measurements are correlated with the process and/or component measurements. High correlation, while not providing direct evidence of the causation of latency, is nevertheless used to identify likely factors (i.e., processes, components) contributing to latency. The latency measurements are plotted over time, the plot including supplemental information indicating, at any time instant, likely factors contributing to the storage system latency.