Neural Network Load Balancing for Integration Processes

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current integration processes lack efficiency in executing triggered events due to preset schedules and inability to analyze metadata in real-time, leading to unnecessary executions and suboptimal resource allocation in cloud computing nodes.

Innovation Solution

The implementation of an enhanced connector SDK with a trained triggering event correlating neural network for real-time analysis of metadata and an execution location optimizing neural network to determine optimal execution timing and location for triggered integration processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If integration processes are executed based on preset schedules, then system simplicity is maintained, but execution efficiency deteriorates due to unnecessary executions

Engineering Contradiction:

Improveexecution efficiencyVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system performs preliminary analysis of metadata before triggering integration process executions. The neural network evaluates incoming data changes in advance to determine whether they warrant process execution, preventing unnecessary executions before they occur. This preliminary filtering action resolves the contradiction by maintaining simple scheduling while improving execution efficiency through intelligent pre-assessment.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements feedback mechanisms where execution results and metadata analysis outcomes are fed back into the neural network model. This feedback loop allows the system to learn from past executions and refine its triggering decisions, improving execution efficiency while maintaining manageable system complexity through adaptive optimization rather than rigid complex rules.

Inventive Principle:

Principle #23Feedback

2Reliability

If integration processes are triggered by all metadata changes, then responsiveness is improved, but resource allocation deteriorates due to suboptimal node selection

Engineering Contradiction:

ImproveresponsivenessVSAvoidresource utilization

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system applies local quality by matching specific types of metadata changes to specific integration processes based on their relevance. The neural network analyzes the local characteristics of each data change and determines whether it warrants process execution, rather than applying a uniform triggering rule. This selective approach maintains responsiveness to meaningful changes while reducing resource consumption by avoiding unnecessary executions.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system employs dynamic node selection where the optimal cloud computing node is determined at runtime based on current system state and process requirements. The neural network dynamically evaluates which node is best suited for each integration process execution, allowing the system to respond flexibly to changing conditions and optimize resource utilization while maintaining high responsiveness.

Inventive Principle:

Principle #15Dynamics

3Measurement precision

If real-time metadata analysis is implemented, then execution precision is improved, but processing time increases

Engineering Contradiction:

Improvetriggering accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system replaces traditional mechanical rule-based metadata analysis with a neural network model. This substitution enables more precise triggering decisions by leveraging the neural network's ability to recognize complex patterns and relationships in metadata, achieving higher triggering accuracy without the rigid processing overhead of exhaustive rule evaluation. The neural network processes metadata more efficiently by learning optimal analysis paths.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system optimizes processing time by dynamically adjusting analysis parameters such as the depth of metadata inspection and the complexity of neural network evaluation. Based on the type of metadata change and system conditions, the system adapts its analysis intensity, maintaining high triggering accuracy for critical changes while reducing processing time for routine changes. This parameter adjustment resolves the contradiction between precision and speed.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11500704B2System and method for intelligent real-time listening and load balancing of integration process executions

Publication Date: 2022.11.15 BOOMI LP

AI summary

An information handling system operating an intelligent real time listen and load balance system comprising a processor training a triggering event correlating neural network to identify a correlation between changes made to a dataset during previous triggering events and previous executions of a triggered integration process, based on previous co-occurrences of the triggering event dataset changes and the triggered integration process executions, determining that current changes to the dataset during a current triggering event correlates to the triggered integration process, indicating new or modified data requires execution of the triggered integration process, and determining predicted triggered integration process execution metrics for a plurality of cloud computing nodes based on received performance metrics for the plurality of cloud computing nodes. The processor may also identify an optimal cloud computing node by comparing the predicted execution metrics for each of the cloud computing nodes for execution of the triggered integration process.