AI-Enabled Network Cell Auto-Healing for Faster Failure Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network failures are difficult to detect and resolve quickly due to incomplete or inaccurate network information, lack of observability, and challenges in deploying effective solutions, leading to prolonged downtime and negative impacts on user experience and revenue.
Innovation Solution
An AI-enabled network auto-healing framework that utilizes machine learning to construct knowledge graphs and embedding models from historic records, enabling real-time auto-heal recommendations based on time series data to promptly address network failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional network monitoring methods are used, then network failures can be detected, but the root cause detection is delayed and resolution time is prolonged due to incomplete network information and lack of observability
Solution Approach 1:
The system performs preliminary actions by constructing knowledge graphs and embedding models from historic network records before failures occur. This pre-processing of network data enables the system to have ready-to-use knowledge representations that can be quickly queried and applied when failures happen, reducing both detection and resolution time while maintaining information completeness.
Solution Approach 2:
The patent introduces knowledge graphs and embedding models as intermediary structures between raw network data and failure analysis. These intermediaries transform incomplete or inaccurate network information into structured, queryable knowledge representations that enhance observability and enable rapid root cause identification without losing critical information.
2Productivity
If manual analysis of network failures is performed, then detailed investigation can be conducted, but the process is slow and labor-intensive, leading to prolonged downtime
Solution Approach 1:
The system implements self-service by automatically performing failure detection, root cause analysis, and resolution recommendation without requiring manual intervention. The AI-enabled framework queries the knowledge graphs and embedding models to autonomously identify failures and suggest fixes, dramatically improving productivity while the automated nature offsets the initial system complexity.
Solution Approach 2:
The patent replaces manual mechanical analysis processes with AI-based automated systems. Instead of human operators manually investigating network failures, the system uses machine learning models and knowledge graphs to automatically detect, analyze, and resolve failures, increasing productivity despite the complexity of the AI infrastructure.
3Reliability
If comprehensive network monitoring is implemented, then more network information becomes available, but the complexity of deploying and managing the monitoring system increases
Solution Approach 1:
The monitoring system is segmented into modular components: knowledge graph construction modules, embedding model training modules, failure detection modules, and resolution recommendation modules. This segmentation allows the system to achieve comprehensive monitoring and high reliability while managing complexity through independent, manageable components that can be deployed and maintained separately.
Data Source
AI summary
The present teaching relates to AI-enabled auto-heal of network cells. Bundled embedding models are obtained, via machine learning, based on historic records representing knowledge on past dynamics of a network. Each of the bundled embedding models captures a respective aspect of the past network dynamics. When temporal data is received with real time observations of the network operation, metrics on the performance thereof, and a point of failure, embeddings of the temporal data relating to the point of failure are derived, based on the bundled embedding models, and used to generate, by time series forecasting, a recommendation on an auto-heal resolution. When performance information of the network associated with the point of failure is received, it is used for online learning of learnable parameters associated with the time series forecasting.