Remedial Action Recommendation Model for Network Event Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network event management tools struggle to adequately analyze, respond to, predict, or prevent undesired network events due to the sheer number of events and the interconnected nature of network topologies.
Innovation Solution
A computer program product that includes executable code to receive source alarms and target remedial actions, extract features, process them through a remedial action recommendation model trained on previous data, and produce ranked recommended remedial actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional network event management tools are used to monitor and analyze events, then system monitoring coverage is maintained, but the tools become unable to adequately analyze, respond to, predict, or prevent undesired network events due to the sheer number of events and interconnected nature of network topologies
Solution Approach 1:
The patent segments the complex network event analysis problem into multiple layers of small world networks. Events are organized into hierarchical clusters where similar events are grouped together, reducing the complexity of analyzing the sheer number of events. This segmentation allows the system to manage and analyze large volumes of events by breaking them down into manageable clusters with representative events.
Solution Approach 2:
The patent introduces small world networks as an intermediary structure between raw events and analysis outcomes. These networks serve as a mediator that captures causal relationships and event patterns, enabling the system to handle the interconnected nature of network topologies. The small world networks act as a buffer layer that transforms complex event data into structured relationships that can be efficiently analyzed.
2Measurement precision
If all possible relationships between multiple events are examined to determine causal associations, then event analysis accuracy is improved, but the time required to analyze events becomes excessive and unmanageable
Solution Approach 1:
The patent performs preliminary actions by pre-processing events into small world networks and hierarchical clusters before actual analysis is needed. Causal relationships are pre-identified and structured in the network topology, so when events occur, the system can quickly query pre-computed relationships rather than analyzing all possible event pairs from scratch. This preliminary structuring significantly reduces analysis time while maintaining accuracy.
Solution Approach 2:
The patent implements dynamic event clustering where the granularity and scope of event relationships examined adapts based on the specific analysis context. The system can dynamically adjust which causal relationships to explore in depth versus which to summarize, allowing flexible trade-offs between analysis accuracy and time consumption based on operational needs.
3Productivity
If automated remedial action recommendation systems are implemented, then response efficiency to network events is improved, but the complexity of the system increases
Solution Approach 1:
The patent implements self-service capabilities where the system automatically learns from historical event data and remediation outcomes to improve its own recommendation quality over time. The machine learning components enable the system to autonomously adapt to new event patterns and refine causal relationship models without requiring manual reconfiguration, offsetting the initial complexity increase with long-term operational simplicity.
Solution Approach 2:
The patent utilizes parameter changes in the form of machine learning model parameters that are dynamically adjusted based on incoming data. The system monitors performance metrics and feedback to continuously optimize its recommendation algorithms, allowing it to adapt to changing network conditions and event patterns. This dynamic parameter adjustment enables the system to maintain high productivity while managing complexity through data-driven optimization rather than hard-coded rules.
Data Source
AI summary
Described systems and techniques determine causal associations between events that occur within an information technology landscape. Individual situations that are likely to represent active occurrences requiring a response may be identified as causal event clusters, without requiring manual tuning to determine cluster boundaries. Consequently, it is possible to identify root causes, analyze effects, predict future events, and prevent undesired outcomes, even in complicated, dispersed, interconnected systems.