Proactive Monitoring
Each box in the network may contain lots of Intelligent managed objects.
- One for each line controller
- Intelligent modems
- management of switch/router/hub/server itself.
Program in router called RMON monitors any parameters in MIB high/low/threshold and generates alarms.
Issues about:
- polling - high network traffic, high resilience
- alarms - processing overhead (CPU cycles) in switch, instantaneous (don't have to wait for next poll)
Filtering of alarms
- in network vs. centralised? (FTP is more efficient than SNMP)
- Expert system looks for root cause.
- Proactive Monitoring
Expert Systems
- Automated Troubleshooting Assistance
- Impact Determination
- Hot spot analysis, Integrated Network Planning
- Alarm Correlation and Root Cause Analysis
Customer Communication
CTI
- On incoming calls
- On outgoing calls
Other Issues:
- Expert System requires rules to be written for each custom network when it is designed?
- How do we display faults: scrolling alarms vs. maps
- How to visulise network
- Can we proactivly monitor high level applications as opposed to line up/down.