Network Analytics Engine Graph for Automated Troubleshooting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network troubleshooting approaches require scripting expertise, are limited in the number of entities that can be tracked, and often overlook relationships between monitored metrics, leading to incomplete root cause analysis.
Innovation Solution
A network analytics engine (NAE) uses data mining and natural language processing to extract relationships between networking entities, creating a graph that dynamically enables and disables scripts to monitor key entities, enabling automated debugging and root cause analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If custom software is installed on network devices to track parameters and generate alerts, then troubleshooting automation is improved, but device complexity and installation requirements increase
Solution Approach 1:
The patent introduces an intermediary system that collects network data through standard protocols without requiring custom software installation on network devices. This intermediary layer processes and analyzes the data, enabling automated troubleshooting while avoiding direct modification of network device complexity
Solution Approach 2:
The patent replaces the mechanical approach of installing custom software agents on network devices with a data-driven approach using machine learning models that process standard network data streams, thereby eliminating the need for additional software installation while maintaining automation capabilities
2Adaptability or versatility
If users write scripts to track networking parameters, then customization and monitoring capability are improved, but scripting expertise requirement increases
Solution Approach 1:
The system performs self-service by automatically selecting and configuring the monitoring parameters and scripts needed based on the network topology and historical data, eliminating the need for users to manually write scripts while maintaining customization capabilities
Solution Approach 2:
The patent dynamically changes monitoring parameters based on network conditions and historical analysis, allowing the system to adapt to different scenarios without requiring users to manually configure scripts or understand scripting languages
3Device complexity
If a limited number of entities are tracked in current troubleshooting tools, then system complexity is reduced, but measurement precision and root cause analysis completeness deteriorate
Solution Approach 1:
The patent implements dynamic entity tracking where the system automatically adjusts which network entities to monitor based on current network conditions, historical data, and detected anomalies, allowing flexible scaling from limited to extensive tracking without fixed system complexity constraints
Solution Approach 2:
The patent adds a temporal dimension to entity tracking by analyzing historical network data alongside current states, enabling comprehensive root cause analysis across multiple time points and entities without proportionally increasing system complexity
4Ease of operation
If relationships between monitored metrics are overlooked, then monitoring simplicity is maintained, but troubleshooting effectiveness and diagnostics capability worsen
Solution Approach 1:
The system implements feedback loops that continuously analyze relationships between monitored metrics and use this information to improve future monitoring and troubleshooting, maintaining simplicity while enhancing effectiveness through learned correlations
Solution Approach 2:
The patent performs preliminary analysis of metric relationships during system setup and historical data processing, pre-identifying correlations and dependencies that will be used during troubleshooting operations, thereby maintaining operational simplicity while improving diagnostic effectiveness
Data Source
AI summary
Some implementations of the disclosure are directed to extracting text about networking entities from textual networking data sources, the textual networking data sources including human readable textual data sources; using natural language processing to tokenize the extracted data to obtain tagged data including networking entities and intents; and filtering out from the extracted data, using at least a dictionary of networking terms, data unrelated to networking troubleshooting. Further implementations are directed to building a graph using at least the tagged data including the networking entities and intents, where the graph includes a plurality of nodes and a plurality of links, where each of the plurality of nodes includes a networking entity.