Real-Time Streaming Analytics for Telecom Network Issue Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Telecommunication network operators face challenges in analyzing vast amounts of data to identify and address issues affecting customer experience, leading to delayed issue resolution and increased customer dissatisfaction due to the complexity and cost of current data analysis processes.

Innovation Solution

The system, referred to as SIRCA, employs real-time streaming analytics and machine learning to extract key performance indicators, identify issues, and provide root cause analysis, using a social network-style interface for accessible insights and self-learning thresholds for outlier detection, enabling timely and efficient issue identification and resolution without requiring specialist involvement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If real-time streaming analytics is implemented to quickly identify network issues, then the response time and customer satisfaction improve, but the system complexity and computational resources required increase

Engineering Contradiction:
Improveissue resolution timeVSAvoidsystem complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The system segments the complex data analysis task into multiple components: data collection from network elements, real-time streaming processing, machine learning-based issue identification, and root cause analysis. This segmentation allows each component to be optimized independently while maintaining overall real-time performance without requiring complete system redesign

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary streaming analytics platform that sits between raw network data and decision-making systems. This intermediary layer processes and analyzes data in real-time, transforming complex network data into actionable insights without requiring direct complex processing in the core network elements or decision systems

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If specialized experts are involved in analyzing network data to ensure accurate issue identification, then the measurement precision improves, but the operational efficiency and response speed decrease

Engineering Contradiction:
Improveissue identification accuracyVSAvoidoperational efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The system implements self-service capabilities through machine learning models that automatically learn from historical network data and independently identify issues without requiring expert intervention. The system performs self-diagnosis, root cause analysis, and anomaly detection, freeing experts from routine analysis tasks while maintaining high accuracy through continuous learning and adaptation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent incorporates feedback mechanisms where the system continuously monitors network performance, compares actual results with predicted patterns, and uses this feedback to refine its analysis models. This closed-loop approach enables the system to maintain high identification accuracy automatically, reducing reliance on expert oversight while improving operational efficiency

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3436951B1Systems and methods for measuring effective customer impact of network problems in real-time using streaming analytics
Publication Date: 2024.03.20 ANRITSU CO
  • EP3436951B1 patent drawingFigure 1
  • EP3436951B1 patent drawingFigure 2
  • EP3436951B1 patent drawingFigure 3

AI summary

A system used for identifying issues within a telecom network. Data is obtained from sources including probes and network elements. KPIs are identified for real-time streaming aggregation. Streaming data related to the KPIs is aggregated and an approximation of count-distinct subscribers and volume count is calculated, as well as count-distinct subscribers aggregating by each identified KPI. Drill objects found in the aggregated data are identified based on the calculations and real-time trending records are generated and stored for each drill object using an exponential moving average. Baseline averages are generated based on the real-time trending records. An increase in errors can then be detected based on the baseline averages and additionally aggregated real-time streaming data. Deviations in each drill object contributing to the detected increase in errors are then analyzed and a full case report is generated based on details of the deviations.