Live Kernel Dump Prioritization for Online Machine Issue Resolution

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Cloud computing environments face issues with physical computing devices leading to virtual machine unavailability and extended downtime, causing service outages and revenue loss due to conventional offline processing of live kernel dumps, which are time-consuming and do not provide real-time resolution.

Innovation Solution

The method prioritizes processing and transmission of live kernel dumps by flagging them for expedited analysis, compressing and encoding, and using a prioritized queue to enable near-real-time issue resolution, reducing downtime by implementing least disruptive remedial actions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional offline processing of live kernel dumps is used, then analysis completeness is improved, but virtual machine downtime increases from 3 minutes to 26 minutes

Engineering Contradiction:
Improveanalysis completenessVSAvoidvirtual machine downtime
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by initiating kernel dump collection and prioritized processing before the virtual machine is fully taken offline. The flagged kernel dump queue pre-prioritizes and prepares analysis workloads, enabling faster transition to online resolution mode and reducing overall downtime while maintaining analysis completeness.

Inventive Principle:
Principle #10Preliminary action

2Speed

If live kernel dumps are processed with high priority, then processing speed is improved, but system resource contention increases

Engineering Contradiction:
Improveprocessing speedVSAvoidsystem resource contention
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The system segments the kernel dump processing workload by creating a dedicated flagged kernel dump queue separate from the standard kernel dump queue. This segmentation isolates high-priority analysis tasks from routine processing, enabling faster processing speeds for critical issues while preventing resource contention from affecting overall system stability.

Inventive Principle:
Principle #1Segmentation

3Difficulty of detecting and measuring

If virtual machines are moved to another physical computing device for diagnosis, then issue isolation is improved, but service continuity is disrupted

Engineering Contradiction:
Improveissue isolationVSAvoidservice continuity
Core Design Contradiction:
Difficulty of detecting and measuringVSReliability

Solution Approach 1:

The system introduces an intermediary approach by using targeted kernel dump analysis to identify and isolate specific issues without requiring physical migration of virtual machines. The flagged kernel dump queue acts as an intermediary mechanism that enables problem diagnosis and resolution while keeping virtual machines in their original hosting environment, thereby maintaining service continuity.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4645088A1System and method for online machine issue resolution using live kernel dump file analysis
Publication Date: 2025.11.05 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4645088A1 patent drawingFigure 1
  • EP4645088A1 patent drawingFigure 2
  • EP4645088A1 patent drawingFigure 3

AI summary

A method, computer program product, and computing system for detecting, using an agent, an issue associated with a target computing device. A kernel dump file is obtained by initiating a kernel dump concerning the issue from the target computing device to the agent. A flagged kernel dump file is generated by flagging the kernel dump file for prioritized processing on an agent processing queue. The flagged kernel dump file is queued in a prioritized kernel dump queue. The flagged kernel dump file is uploaded from the prioritized kernel dump queue to an analysis server for online resolution.