Automated Problem Graph Identification in IT Infrastructure Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current IT infrastructure management relies heavily on manual domain knowledge for identifying problem graphs, which is time-consuming and inefficient, as operators need to manually inspect network topology to diagnose root cause problems, increasing the Mean Time to Know (MTTK).
Innovation Solution
A computer-implemented method that analyzes historical timeseries data to infer problem graphs by identifying directional rules and causality relationships, using association rule mining algorithms to automatically detect and predict root cause problems without the need for heuristic rules, thereby reducing operator dependency and accelerating incident resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If operators manually identify problem graphs using domain knowledge, then they can diagnose root cause problems, but the Mean Time to Know (MTTK) increases and the process becomes time-consuming
Solution Approach 1:
The patent replaces the manual mechanical process of operator inspection with an automated computational system. The system uses graph analysis algorithms to automatically identify problem graphs by analyzing dependencies between IT resources, substituting human operators with automated software that processes topology data and alert information to diagnose root causes faster.
Solution Approach 2:
The system enables self-service diagnosis by automatically generating problem graphs and identifying root causes without requiring operator intervention. The automated system serves itself by collecting data, analyzing dependencies, and producing diagnostic results independently, reducing reliance on manual domain knowledge while maintaining diagnostic accuracy.
2Measurement precision
If operators manually inspect network topology to diagnose problems, then they can identify root cause problems, but the complexity of the process increases
Solution Approach 1:
The patent segments the complex diagnosis process into distinct automated steps: collecting topology data, identifying dependencies between resources, detecting problem states, analyzing graph structures, and determining root causes. This segmentation breaks down the overwhelming manual inspection task into manageable automated components, reducing perceived complexity while maintaining diagnostic precision.
Solution Approach 2:
The system introduces an intermediary automated analysis layer between the raw network data and the final diagnosis. This intermediary process automatically processes topology information, identifies dependencies, and generates problem graphs, serving as a mediator that simplifies the complex relationship between numerous IT resources and their interdependencies without requiring operators to manually navigate the complexity.
3Reliability
If operators rely on manual domain knowledge, then they can diagnose incidents, but the dependency on operator expertise increases and efficiency decreases
Solution Approach 1:
The patent replaces the mechanical reliance on operator domain knowledge with an automated system that encodes diagnostic logic in software. The system substitutes human expertise with algorithmic analysis of topology data and alert patterns, maintaining reliable diagnosis through consistent application of analysis rules while dramatically improving productivity by eliminating manual inspection bottlenecks.
Solution Approach 2:
The system changes the parameter of knowledge representation from human operator expertise to structured data analysis parameters. Instead of relying on operators' internalized domain knowledge, the system transforms diagnostic criteria into explicit computational parameters that can be automatically applied to topology data, ensuring consistent reliable diagnosis while scaling efficiency across multiple incidents simultaneously.
Data Source
AI summary
Method and system are provided for identifying problem graphs in an information technology infrastructure network. The method includes: selecting a set of nodes in a topology, wherein the nodes of the topology represent resources in an information technology infrastructure network; querying historical timeseries data for the selected nodes, wherein the timeseries data records changes in state of the resources represented by the nodes; and analyzing the timeseries data for the selected nodes across problem time periods to produce a set of historical directional rules defining one or more historic problem graph. The method may search the historic problem graphs for current problem state nodes to determine likely causing nodes and affected nodes of a current problem.