Using the results of those probes, Nodemap produces visualizations to convey the "holistic" state of the network. A "drill-down" user interface is provided which permits detailed link-by-link status information to be presented at low-levels, with abstracted summarizations available at higher levels. This enables network operations staff to determine the current state of a network with a single glance, without regard to topology complications or the number of devices on the network.
Using Nodemap, network support staff can be armed with enough information to diagnose (or dismiss!) performance complaints without needing to know every last detail about the way the network is set up. Nodemap is also useful for tracking the flow of DoS packets through complex networks.
Nodemap is also a useful documentation resource: By providing a single place to hold all the topology data for a network, together with a mechanism for automatically producing network diagrams, Nodemap is a valuable resource for keeping track of network infrastructure on an ongoing basis. If updating Nodemap's config file is allowed to become a standard part of an organization's procedures for adding new network resources, the Nodemap config file becomes a one-stop-shop for documenting which bits of the network are connected to which other bits.
Nodemap is scalable. Internode is currently using it to probe hundreds of interfaces on over 100 devices on a national network which spans the Australian continent. Each SNMP run takes approximately 12 seconds, and each "ping" run (for latency and packet loss measurement) takes about 35 seconds. Extremely large networks can be interrogated by a single Nodemap monitoring station within the five minute refresh interval currently used by the software; Even larger networks can be configured if they are heirarchially segmented in the configuration file and probed in parallel.
In furtherance of the "Shameless Promotion" point, we would appreciate it if you could refer to this software as "Internode Nodemap", rather than just "Nodemap", in any written appearances on web sites, documentation, etc, unless there's some particular reason not to (e.g., you're one of our competitors, or you think we're idiots, or whatever). We obviously can't force you to do that, but if there's no reason not to we'd like it if you did anyway. The more obviously we see the aims above satisfied, the more likely we are to release additional software later on.
Download the code. Read the Installation Instructions. And read the rest of the manual to find out how to configure it. There are also some notes regarding installation on Solaris or on other platforms with GD libraries which don't support GIF.
UPDATE: 10 August 2004
Internode Nodemap 1.1 is available (updated 10 August 2004). This is now the new production release. The old nodemap is still available as nodemap-1.0.tar.gz for historical/hysterical reasons.
Configuration examples are provided in the distribution. You'll probably find the config file syntax pretty simple to work with (I certainly do, but I wrote it so I'm probably not the best judge :-)
There are also some screenshots. It's visualization software, so it'd be pretty odd if I didn't show you any of them :-)
There's an Internode Nodemap Mailing List. If you make any software fixes, please send the list some email so we can incorporate them into the next version. Note that this software is licensed under the GPL, so if you make any changes to the software you're not permitted to redistribute them to third parties unless you provide the source (but you can make as many private alterations as you wish without obligation).
As changes, bugfixes, new features, etc are integrated into the code, they'll show up in the current development version, which you can download if you dare.
Internode's network footprint extends to all capital cities in Australia and many regional areas throughout South Australia. We have staff in Adelaide and Sydney. Our NOC is in Adelaide.
I wrote Nodemap because it's impossible to get a global view of an entire network by looking at MRTG graphs and Netflow logs. This software scratched my itch. Hopefully it'll be able to scratch yours too.
Mark Newton, May 2004