Node Load Coefficients for Multi-Threaded Bottleneck Identification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In multi-threaded multi-path computing environments, identifying the root cause of high transaction response times is challenging due to the complexity of interactions between multiple nodes, where each node can contribute to bottlenecks, making it difficult for APM software to determine the exact cause of delays.

Innovation Solution

A method and system that analyze interaction data to calculate node load coefficients by determining the ratio of inbound and outbound requests across nodes, identifying which nodes are causing bottlenecks by analyzing timestamped data and determining queuing issues, thereby prioritizing nodes for service.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If APM software collects and analyzes large amounts of interaction data from multiple nodes in multi-threaded multi-path computing systems, then the ability to identify bottlenecks and determine root causes of performance issues is improved, but the complexity of analyzing and processing this data increases significantly

Engineering Contradiction:

Improvebottleneck identification accuracyVSAvoiddata analysis complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the complex analysis task by introducing path-specific metrics that break down the overall system performance into individual path contributions. Each path's load coefficient is calculated separately, allowing the system to identify which specific paths are causing bottlenecks without analyzing all interaction data uniformly.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces path load coefficients as intermediary metrics that mediate between raw interaction data and bottleneck identification. These coefficients serve as intermediate calculations that simplify the analysis by providing a standardized measure of path utilization, making the complex multi-node data more manageable.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If APM software attempts to identify the root cause of high transaction response times by analyzing all nodes, then the completeness of problem diagnosis is improved, but the time and computational resources required increase

Engineering Contradiction:

Improvediagnosis completenessVSAvoidanalysis time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent applies local quality by focusing analysis on specific paths that exhibit abnormal load coefficients rather than uniformly analyzing all nodes. By identifying paths with load coefficients exceeding thresholds, the system concentrates diagnostic efforts on localized problem areas, reducing overall analysis time while maintaining diagnostic completeness.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent uses partial action by calculating path load coefficients only for paths that are part of problematic transactions rather than all possible paths. This selective approach reduces computational overhead while still identifying the root causes of performance issues.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10042733B2Determining where bottlenecks occur in multi-threaded multi-path computing systems

Publication Date: 2018.08.07 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

One or more processors analyze interaction data for two or more nodes within a plurality of nodes processing a computing transaction. One or more processors determine a number of inbound requests and a number of outbound requests for the two or more nodes within the plurality of nodes. One or more processors determine whether one or more nodes within the plurality of nodes are limiting computing performance.