GNMS - Gnome Network Management System

About FAQ News Gallery Download Get involved

---

Introduction

Beginning...

Early in 2003, I was surprised to not found a free NMS tool on Linux, a lightweight tool version of HP OpenView, Castle Rock SNMPc, Ipswitch What's Up or just a simple network supervisor. I knew Cheops-NG but it was no use with what I sought.
Thus I started working on my own, it was on mid January 2003.

GNMS is a free Network Management System, a tool to monitor state of network elements.
It is written, for Linux platform, in Ruby with Gnome2 bindings.

What is a NMS ?

A NMS, stands for Network Management System, is a system used to managed the state of network elements.
States can be determined though two ways:

• Semi-real time by polling informations
• Real time by catching SNMP trap events and SYSLOG messages

Main features

Visual representation

Elements monitored (aka nodes) are displayed and mapped on the main window.

• Each node icon represents it's actual OS vendor logo or device type (also defined by the node form)
  Device type are as follows:
  

  Switch		Router		Firewall
  Device		Printer		Host
  Server		Gateway		Network

• Each node color represents it's actual states among



CRITICAL	All metric we monitor are not responding
MAJOR	Almost all metric we monitor are not responding
WARNING	At least one metric we monitor is not responding
MINOR	At least one metric we monitor is not responding
NORMAL	Node is ok
UNMANAGED	Node is not monitored
UNKNOWN	Node status is not yet determined


This state is computed from four monitoring values: Service, SNMP, Custom script, Trap.

Semi-real time to real time monitoring

Active monitoring via polling:

• Node availability is done via ICMP queries
• Service availability is done via NMAP command wrapper
• Threshold exceeded is done via SNMP queries
• Other metric can be followed via Custom Monitoring plugins for:
  url monitoring, database monitoring, application monitoring, traffic monitoring, vendor specific monitoring, resource monitoring, service monitoring

Passive monitoring via received triggers:

• SNMP trap server
SNMPv1 or SNMPv2 traps are received and processed to identify if trap source address is known and if it is configured to handle this OID.

• SYSLOG server messages
UDP messages are received and processed to identify if message source address is known and if it is configured to handle this message defined by a regular expression.

These services can be enabled/disabled and configured on demand.

Notification

Alerting can be done via one of the following method:

• Syslog
• Mail
• Popup
• Sound (just a beep or sound file playing)
• SMS* through Ruby-SMS lib embedded but needs Ruby-Serialport lib
• External script execution

Storage

Configuration file is stored as plain text YAML file
Network nodes topology can be stored as:

• XML file
• SQL DB*

Misc.

• Auto discovery: IP address, MAC address, OS (by using NMAP and xprobe2 finger printing)
• Network topology*
• Advanced node finder
• IP calculator
• MIB browser*
• Tree node manager
• Windowed/fullscreen node map
• Plugins support

* still need some work

---

© 2003-2007, David Maciejak