Learning-Based IT Problem Management for Proactive Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current IT systems lack an automated and proactive approach to identify and manage problems across enterprise computing systems, leading to increased complexity, service disruptions, and inefficiencies in problem detection and resolution, particularly due to the absence of a global view, correlation of metrics, and reliance on manual analysis and reactive incident-based workflows.
Innovation Solution
A learning-based method and system that utilizes machine learning techniques to analyze multi-dimensional data from enterprise resources, including metrics and incidents, to detect problems using propositional logic, refine detection rules, and generate feedback to suppress unwanted issues, thereby providing proactive problem management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual analysis is used to identify problems, then flexibility and adaptability are maintained, but productivity and detection accuracy are reduced
Solution Approach 1:
The system performs automated self-analysis of IT operations data using machine learning models to detect problems and generate problem requests, eliminating the need for manual analysis while maintaining adaptability through learnable patterns from historical data
Solution Approach 2:
Manual mechanical analysis processes are replaced with automated machine learning-based detection systems that process metrics and incident data to identify problems, significantly improving detection accuracy and productivity
2Device complexity
If reactive incident-based workflows are used, then simplicity is maintained, but reliability and service continuity are worsened
Solution Approach 1:
The system proactively detects potential problems by analyzing metrics and incident patterns before they escalate into full incidents, enabling preventive action to be taken and maintaining service continuity while keeping workflows manageable
3Productivity
If automated problem detection is implemented, then productivity is improved, but device complexity and implementation difficulty increase
Solution Approach 1:
The machine learning model serves multiple functions: detecting problems, generating problem requests, identifying patterns, and learning from feedback, consolidating multiple detection mechanisms into a single universal system that improves productivity without proportionally increasing complexity
4Measurement precision
If comprehensive enterprise data is analyzed, then measurement precision and problem detection accuracy are improved, but loss of time and processing overhead increase
Solution Approach 1:
The system analyzes comprehensive enterprise data but focuses detection efforts on critical patterns and high-impact metrics identified through machine learning, achieving high detection accuracy without processing every data point equally, thus reducing overall processing time
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
An architecture of an enterprise can be fully on on-premises with multiple computing systems or on cloud or on Hybrid. Enterprise information technology (IT) departments are continually adopting new technologies and changes leads more complex, and increased probability of service disruption, through malfunctions in software, hardware, networks, natural disasters, simple human error. Embodiments of the present disclosure provide a method and system to detect problem associated with the computing system in an enterprise. A plurality of data associated with each resource in the enterprise is received. The plurality of data is considered to iteratively perform (a) derive a parameter associated with a metric data, and an incident data to obtain an analyzed data, (b) detect a problem associated with each resources by processing the plurality of analyzed data based on a propositional logic, and (c) generate a feedback associated with each problem of the resources in a subsequent iteration.