Networking Kid / Addresses, Names And Routes

Three Words People Use As If They Were One

A name, an address and a path are different things doing different jobs.

A name is a label for people. It is chosen to be memorable and pronounceable, it can be sold, moved and pointed somewhere else at will, and it says nothing whatsoever about location. An address identifies a particular network attachment point. It is structured so that equipment can sort it efficiently, it usually reflects something about where in the network that attachment sits, and it can change when the attachment moves or is reissued. A path is the actual sequence of handovers a message takes between two addresses on a given day, and it is not a property of either end. It is a property of the network at that moment.

Confusing these three produces most beginner misunderstandings, and the confusion is understandable because everyday speech runs them together. People assume a name is a place, so they cannot explain how a site keeps working after moving to entirely different equipment on another continent. People assume an address is fixed and personal, so they are puzzled when a device is given a different one after being switched off overnight. People assume a path is stable, so they are surprised that the same request can be quick in the morning and slow at night with nothing at either end having changed at all.

Hold them apart and a great deal becomes obvious. A name points at an address. An address can be reached by many possible paths. The pointing can be changed without touching a single piece of equipment. The path can change without anybody changing anything at all, anywhere. Each of the three can be perfectly correct while another is wrong, which is exactly why a fault can look mysterious until you stop and ask which of the three you are actually looking at. Half of clear thinking about networks is just refusing to let these words blur together.