Virtual Machine Failure Diagnosis via Multi-Dimensional Service Interaction Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing virtual machine failures in cloud systems rely on operation logs, which are insufficient for tracing service interaction information and correlating multiple events, making it difficult for maintenance personnel to determine the cause of failures in a comprehensive manner.
Innovation Solution
An auxiliary method and system that generates graphic multi-dimensional display information from running information of virtual machines, including state, operation, alarm, and CPU usage data, to facilitate correlation analysis and root cause identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If operation logs are used for failure diagnosis, then the system structure remains simple, but the ability to trace service interaction information and perform correlation analysis deteriorates
Solution Approach 1:
The patent transforms flat operation logs into multi-dimensional service interaction graphs by adding temporal, causal, and hierarchical dimensions. This allows maintenance personnel to view service interactions from multiple perspectives (time sequence, cause-effect relationships, service hierarchy) simultaneously, enabling comprehensive correlation analysis while maintaining system structure simplicity.
Solution Approach 2:
The patent introduces a service interaction graph as an intermediary layer between operation logs and failure diagnosis. This intermediary structure processes and organizes raw log data into meaningful service interaction patterns, preserving complete service interaction information while keeping the underlying system structure simple and manageable.
2Ease of manufacture
If operation logs are used for failure diagnosis, then the implementation remains simple, but the accuracy and comprehensiveness of failure cause determination deteriorates
Solution Approach 1:
The patent performs preliminary organization and structuring of service interaction information before failure diagnosis occurs. By pre-building service interaction graphs that capture temporal and causal relationships, the system prepares comprehensive diagnostic data in advance, enabling accurate failure cause determination without complicating the actual diagnosis implementation.
Solution Approach 2:
The patent creates a virtual copy of service interactions in the form of service interaction graphs, which mirror real service relationships but exist as structured, analyzable data models. This copying approach enables comprehensive multi-dimensional analysis of failure causes while keeping the physical implementation simple.
3Quantity of substance
If traditional operation logs are used, then the information volume is manageable, but the ability to perform multi-dimensional correlation analysis deteriorates
Solution Approach 1:
The patent adds multiple analysis dimensions (temporal sequence, causal relationships, service hierarchy, interaction frequency) to the service interaction information without significantly increasing information volume. This dimensional enrichment enables comprehensive correlation analysis while keeping the information manageable through efficient graph-based representation.
Solution Approach 2:
The service interaction graph serves multiple diagnostic functions simultaneously: it traces service interactions, establishes causal relationships, identifies failure patterns, and supports various analysis perspectives. This multi-functionality enables versatile multi-dimensional correlation analysis without proportionally increasing information volume.
Data Source
AI summary
Embodiments of the present invention relate to an auxiliary method, apparatus, and system for diagnosing a failure of a virtual machine. An operation and maintenance system receives a failure message when a virtual machine fails; obtains running information of the virtual machine according to identifier information carried in the message, where the running information is used for diagnosing the failure of the virtual machine; generates graphic multi-dimensional display information of the virtual machine according to the running information; and sends the display information to a user end for displaying. The present invention provides an auxiliary method for diagnosing a failure of a virtual machine. The method may provide correlation analysis for diagnosing a failure of a virtual machine, allow the user end to diagnose the cause of the failure of the virtual machine in a comprehensive multi-dimensional manner, and improve the convenience and accuracy of diagnosing the virtual machine.