ML Network Diagnosis via Device Logs and CDRs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Telecommunications networks face challenges in accurately diagnosing and resolving network issues due to insufficient customer reports, lack of device-specific log data, and a slow reporting process, which hampers timely investigation and corrective actions.
Innovation Solution
A machine learning model is trained using enhanced call detail records (CDRs) supplemented with device-specific information and network data to predict the cause of network issues and perform corrective actions, improving the reporting process and reducing downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional customer reporting methods are used, then network issues can be detected through customer reports, but the reporting process is slow and lacks sufficient information for accurate diagnosis
Solution Approach 1:
The system proactively collects device-specific log data, network data, and call detail records before customers report issues. By preparing diagnostic data in advance and continuously monitoring network parameters, the system reduces the time needed for diagnosis when issues occur, while maintaining high accuracy through comprehensive pre-collected information.
Solution Approach 2:
The machine learning model acts as an intermediary that processes and correlates multiple data sources (customer reports, device logs, network data, CDRs) to generate accurate diagnoses. This intermediary system integrates information from various sources that would otherwise remain separate, enabling fast and accurate diagnosis by automatically correlating data across different levels.
2Loss of information
If customer reports are used as the primary data source, then network issues can be identified, but the reports lack device-specific log data and sufficient information for accurate cause determination
Solution Approach 1:
The system merges multiple data sources including customer reports, device-specific log data, network data, and call detail records into a unified diagnostic framework. By combining these diverse information sources, the system achieves comprehensive information completeness for accurate diagnosis while the machine learning model handles the complexity of integrating and processing this multi-source data automatically.
Solution Approach 2:
The machine learning model serves multiple functions: it processes customer reports, analyzes device logs, correlates network data, and generates diagnoses across different network levels (device, LAN, metropolitan). This multi-functional approach allows a single system to handle diverse data types and diagnostic requirements without requiring separate specialized systems for each function.
3Productivity
If manual fault management processes are used, then network issues can be detected and corrected, but the process requires multiple channels of communication and is time-consuming
Solution Approach 1:
The system implements automated feedback loops where the machine learning model continuously receives new data, generates diagnoses, and triggers appropriate corrective actions. This automated feedback mechanism eliminates the need for multiple manual communication channels by automatically processing information from detection through diagnosis to resolution, significantly improving productivity while maintaining high automation levels.
Solution Approach 2:
The machine learning model performs self-service by automatically analyzing collected data, identifying network issues, determining causes, and initiating corrective actions without requiring manual intervention at each step. This self-service capability allows the system to resolve issues autonomously, improving resolution speed while maintaining extensive automation throughout the fault management process.
Data Source
AI summary
A machine learning system automatically diagnoses and resolves issues in a telecommunications network. When a customer reports a network issue using a mobile application, the device performs a diagnostic test, such as a speed test. In addition, network logs or performance metrics during occurrence of the network issue are collected. The results of the diagnostic are used as inputs to a machine learning model in combination with the network logs or metrics to predict the cause of the network issue or to perform a corrective action.