Virtual Service Session Failure Prediction Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtual computing environments with numerous machines or virtual services, administrators face challenges in tracking the health of these services, leading to potential downtime and persistent session failures that adversely affect user experience, often discovered only after failures occur.
Innovation Solution
A server computer system uses machine learning techniques to proactively predict session failures by analyzing feature values, applying supervised learning models like DNNs and RNNs, and executing corrective operations such as ending sessions or restarting services to prevent black hole states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If administrators manually track the health of virtual services, then they can detect failures, but the complexity and time required increases significantly with the number of services
Solution Approach 1:
The system enables virtual services to self-diagnose their health status by automatically collecting and analyzing their own feature values (CPU usage, memory usage, disk usage, network usage) without requiring administrator intervention. The machine learning model autonomously evaluates these self-reported metrics to predict session failures.
Solution Approach 2:
The patent replaces manual administrative tracking with an automated machine learning-based prediction system. Instead of administrators manually monitoring and detecting failures, the system uses supervised learning models (DNN, CNN, RNN, LSTM) to automatically analyze service metrics and predict failures before they occur.
2Reliability
If administrators discover black hole states after failures occur, then they can respond to issues, but user experience is already adversely affected
Solution Approach 1:
The system performs preliminary actions by predicting session failures before they actually occur. The machine learning model analyzes current service metrics and historical data to identify patterns indicating impending failures, allowing administrators to take corrective action (restart services, allocate resources) before the black hole state is entered, thus preventing user experience degradation.
3Measurement precision
If the system monitors all virtual services continuously, then failure prediction accuracy improves, but computational resources and system complexity increase
Solution Approach 1:
The system applies partial monitoring by selectively focusing computational resources on services that show predictive indicators of failure. Rather than uniformly analyzing all services at maximum depth, the machine learning model identifies and intensively monitors only those services with feature value patterns suggesting impending failures, reducing overall computational overhead while maintaining high prediction accuracy for critical cases.
Data Source
AI summary
A server computer system configured to proactively predict a session failure of a virtual service is provided. The server computer system includes a memory and at least one processor coupled to the memory. The at least one processor is configured to receive one or more feature values associated with the virtual service. The processor can then evaluate a likelihood of session failure of the virtual service, such as a session launch failure, unresponsive state, or persistent session failure, in a future time interval based on the received feature values. The processor can then determine that the likelihood of session failure satisfies a classification test indicating the session failure is likely. Responsive to the determination, the processor can then execute a corrective operation, such as to end a user session, disable the virtual service, restart the virtual service, or render a user notification.