One of the fundamental advantages of a visualization over standard traceroute output is that you can see the actual network topology, not just a list of IP addresses. But, as you might expect, these topologies can become quite complex. In this post, you’ll learn about some easy ways we provide to simplify network topologies. That way, you can focus on finding problems instead of playing connect the dots.
Hiding Intermediate Hops
In our path visualization, ThousandEyes agents (either our public agents or private agents that you deploy) are on the left and the destinations on the right. Each dot (or “node”) represents a different IP address. In the middle of the path, many of these IP addresses will be for routers inside third-party networks, such as tier-1 ISPs. When viewing the whole topology, you might see something like this:
Most of the time, the details of the path through the core of the Internet aren’t of much concern to you. You can simplify the topology by hiding these intermediate hops using the following control:
The result looks something like this:
Those numbers on the dotted lines tell you the number of hidden hops. Of course, for those times when the intermediate nodes matter, you can interactively reveal hidden nodes by adjusting the slider from either end.
Another way to simplify network topologies is with interface grouping. Interface grouping is a way of adding some semantic information to traceroute data. One physical device often has many IP addresses, so what appears as multiple nodes may actually just be different interfaces on the same device.
Interface grouping allows you to specify sets of IPs that belong to a single device, combining them into a single node in the visualization.
The result is a simplified graph that better represents your actual network topology. This is just one small example of what you can do with interface grouping. How much would it simplify your network graph?
Stay tuned for a post about our upcoming auto-grouping feature; leveraging reverse DNS and regular expressions, we’ll be able to group topologies with a single click.