Zed, We Have a Bug

Those of my three readers who know me should know that I really don’t like network maps. That said, we on the OpenNMS team are trying to fix most of the issues I’ve had with the usefulness of maps in network management.

In 1.12 we introduced two new maps. One, the “geographical” map, leverages information in the assets table of the database to display devices on a map based on their location. It was nicely done, with smooth transitions and automatic aggregation of devices so that even with thousands of nodes the map is usable.

The second one is the “topology” map. At first it only shipped with two topologies: one based on Layer 2 connections discovered by linkd, and the other based on VMWare relationships retrieved from vSphere and showing how host machines, guests and storage are all interconnected. Since there is a topology API, new topologies can be added (Juniper is working on one to show logical and physical connections between Juniper devices in a network).

Anyway, David was playing around with a dataset and it produced a pretty cool image. I knew that we probably had bugs in our code, but still …

[Note: one person pointed out that this is more of an arachnid than an insect, to which I have to reply that I use the term “bug” to mean any kind of creepy crawly thing, and thus it is not necessary for bug == insect.]