Discovering Representative Composite CI Patterns in IT Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mining frequent composite configuration item patterns from large, complex Configuration Management Data Base (CMDB) graphs is challenging due to their massive size, repetitive labels, and computational resource constraints, which existing algorithms struggle to address effectively.
Innovation Solution
A method involving data mining, clustering, and network-flow based approaches to extract extended frequent composite CI patterns, using random walks and diffusion-based graph similarity techniques to identify representative patterns, and pruning infrequent patterns, facilitating efficient discovery of meaningful IT infrastructure patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing algorithms are used to mine frequent composite CI patterns from large CMDB graphs, then pattern discovery can be performed, but computational resources are excessively consumed and the process becomes inefficient
Solution Approach 1:
The patent segments the large CMDB graph into smaller subgraphs based on graph similarity metrics. By dividing the massive graph into manageable segments, the algorithm can process each segment independently, significantly reducing computational resource consumption while maintaining the ability to discover frequent composite CI patterns across the entire IT infrastructure
Solution Approach 2:
The patent transforms the graph processing task by changing parameters such as using graph similarity thresholds to filter and prioritize subgraphs for pattern mining. This parameter-based filtering approach allows the system to focus computational resources on the most relevant graph segments, improving productivity while reducing overall resource consumption
2Reliability
If the entire CMDB graph is processed to ensure comprehensive pattern discovery, then all possible patterns can be found, but the processing time and computational complexity increase significantly
Solution Approach 1:
The patent introduces graph similarity as a filtering parameter to identify and process only the most relevant subgraphs. By setting similarity thresholds, the system maintains reliability in pattern discovery by ensuring that representative subgraphs are processed, while significantly reducing processing time by excluding redundant or less relevant graph segments from comprehensive analysis
3Productivity
If graph similarity techniques are used to identify representative subgraphs, then computational burden is reduced, but the complexity of the algorithm increases
Solution Approach 1:
The patent introduces graph similarity metrics as an intermediary mechanism between the raw CMDB graph and the pattern mining process. This intermediary layer filters and prioritizes subgraphs before they enter the pattern discovery pipeline, improving processing efficiency by reducing the input data volume, while the added algorithmic complexity is localized to the similarity computation stage
4Measurement precision
If frequent pattern mining is performed on the complete graph, then all frequent patterns are identified, but the algorithm becomes infeasible for large-scale IT systems
Solution Approach 1:
The patent divides the complete graph into multiple subgraphs based on similarity criteria, allowing frequent pattern mining to be performed on smaller, manageable segments. This segmentation maintains measurement precision for pattern frequency within each segment while making the overall algorithm feasible for large-scale IT systems by avoiding the intractability of processing the complete graph in one operation
Data Source
AI summary
A method for discovering representative composite configuration item (CI) patterns in an IT system that includes a plurality of configuration items may include data mining a graph representing the IT system to extract extended frequent composite CI patterns. The method may also include clustering the extended frequent composite CI patterns into clusters based on similarity between the maximal frequent composite CI patterns. The method may further include extracting a representative composite CI pattern for each of the clusters, and using an output device, outputting the representative composite CI pattern for each of the clusters.