Network Management Systems

1
Network Management Systems

• J.J. Ekstrom
• March 2006

2
Basic Components

• Discovery and Topology Analysis
• Data Management
• Monitoring
• Control
• Fault Notification
• Fault Analysis and Simplification
• Root Cause Analysis
• Dashboard/Extended UI
• Network Design for Management

3
Discovery and Topology Analysis

• Topology Discovery Protocols
• Layer 2 discovery digs deep (Network Computing
  2003)
• 802.1ab, CDP, EDP
• Spanning tree information
• Forwarding table information

4
Data Management

• Proprietary file data (SolarWinds)
• Proprietary Object Store (HP OpenView)
• Relational DB (CA Unicenter, OpenNMS)
• Stores
• Node information and current state
• Link information and loading
• Port information and state

5
Monitoring

• SNMP v1,v2,v3
• Rmon
• HTTP and XML
• Telnet
• Transparent agents in network

6
Control

• SNMP
• HTTP and XML
• Telnet
• Proprietary management protocols

7
Fault Notification

• SNMP Traps
• Monitored rule exceptions (error count too high)
• Proprietary notifications
• Separate agent monitoring rules

8
Fault Analysis

• Manual experienced people under stress!
• Automated
• Expensive solutions
• Need a lot of fine tuning to a network
• Still need experienced people

9
Root Cause Analysis

• You still need to figure out what really
  happened.
• Symptoms (faults) often appear unrelated to
  actual fault.
• Smart People still needed.

10
Dashboard / Extended UI

• Custom to platform (NetView, Solar Winds, CA
  Unicenter, HP Open View)
• Web based (OpenNMS, HP OpenView, HP Procurve
  Manager)

11
Network Design for Management

• Make links match loads
• Monitor design loads against actual
• Monitor error rates (ports at end of links)
• Make sure you can read the data from your
  devices.
• Make sure topology is discoverable and can be
  monitored.