Machine Learning Software Resiliency Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting software application resiliency are inefficient, as they rely on manual review of vast amounts of data, making it difficult to identify potential issues before system outages occur, leading to costly downtime and reputational risks.
Innovation Solution
A computer-implemented method using machine learning techniques to analyze software construction and operation variables, training models to predict error rates and resiliency scores for software applications, providing indications of likelihood and severity of resiliency issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual review of application data is used to detect outages, then system administrators can review data, but the process is extremely time-consuming and impossible for vast amounts of data
Solution Approach 1:
The patent replaces manual mechanical review of application data with an automated machine learning-based detection system. The system uses ML models to automatically analyze application data, predict resiliency issues, and identify outages, eliminating the need for human administrators to manually review vast amounts of data while maintaining reliable outage detection capability.
Solution Approach 2:
The system enables self-service outage detection by automatically analyzing application data and generating predictions about resiliency issues without requiring human intervention. The machine learning models autonomously process data, identify patterns, and alert administrators to potential outages, allowing the system to monitor itself continuously.
2Reliability
If manual review of application data is used, then administrators can identify issues, but system outages occur before any customer complaints are received
Solution Approach 1:
The patent implements preliminary action by using machine learning models to predict resiliency issues before they manifest as actual outages or customer complaints. The system analyzes application data proactively to identify potential problems early, allowing administrators to take preventive measures before failures occur and before customers notice any issues.
Solution Approach 2:
The system establishes continuous feedback loops where application data is constantly monitored, analyzed by machine learning models, and used to generate predictions about potential resiliency issues. This feedback mechanism enables the system to detect patterns and alert administrators to emerging problems before they escalate into outages or customer complaints.
3Reliability
If machine learning models are trained on software construction variables, operation variables, and error rates, then resiliency can be predicted proactively, but the system complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the complex resiliency prediction system into distinct modular components: data collection modules for gathering construction and operation variables, machine learning training modules for model development, prediction modules for generating resiliency scores, and alert modules for notifying administrators. This segmentation makes the complex system more manageable and maintainable while preserving high prediction accuracy.
Data Source
AI summary
Techniques are described for predicting resiliency of software applications. Techniques provide for obtaining one or more software construction variables, software operation variables, and an error rate associated with a first software application. This includes training a machine learning model to predict a resiliency of a particular software application using the software construction variables, the software operation variables, and the error rate for each of the plurality of first software applications. A software construction variable and a software operation variable associated with the second software application are obtained. The trained machine learning model is applied to the software construction variable and the software operation variable associated with the second software application to predict an error rate for the second software application. Then a resiliency for the second software application is determined based upon the predicted error rate and display an indication of the resiliency for the second software application.