Intelligent Big Data Driven Network Management System
1. TYPES OF BDD-NMS
[Offline Model]
- Analysis of Log Data & Proactive Monitoring
- Use of Log Data at Non-realtime Environment
[Realtime Online Model]
- Realtime Data Analysis & Proactive Monitoring by accessing DAS/ICS Networks
- Client/Server Model (HA, Database Replication)
- Multiple Client Support (PC, Web, Mobile, Email, Messenger, etc)
2. FEATURES - Compared to Traditional NMS
Competitiveness
- Based on expert knowledge of repeater equipment & RF system
- Concentrated on Smart Data
- System upgrade thru continuous research of the managed system
Effectiveness
- Significant reduction of maintenance Costs by pre-maintenance by proactive monitoring
- Reduced re-visit rate of disabled site
- Reduced social cost and confusion caused by sudden failure of mobile communication equipment
3. CASE STUDY
Log Data
- Machine Data: ICS (Interference Cancellation System) log file
- Log Data: 30 Months
- Number of Logs: 24,851
- Number of ICS Equipment: 861
Analysis
- Analysis between Alarm/Equip
- Intensive Monitoring
- Alarm Analysis
- Equipment Group Analysis
- Association Rules / Sequences
Proactive Monitoring
- Group Monitoring Algorithm (A, C Group)
- Intensive Monitoring Algorithm
Results
1) Intensive Monitoring Algorithm:
- Re-failure rate of the equipment found by Intensive Monitoring Algorithm is more than 90%.
- To lower the overall failure rate, it is very important to find the intensively monitored equipment.
2) A Group Monitoring Algorithm:
find equipment & alarm type around 70% prediction
3) C Group Monitoring Algorithm:
find equipment & alarm type around 50% prediction
4. UI
