The original Roman sources, and the few commonly available books on the subject, are fairly clear that the Germanic peoples were not town dwellers. They don't seem to have built hill forts or oppida as the Celts/La Tene culture did. What I'm not certain as to is whether the norm was dwelling in agricultural communities (villages) or isolated farmsteads. I've been trying to find an answer to this for months and it's driving me insane, so I'll go ahead and post it here in the hopes that someone can enlighten me.
It depends a bit on region; I will answer here for the Netherlands and Denmark, areas that I am most familiar with and with which the Romans had the most extensive contact.
The true medieval Village had not yet emerged, but clusters of farmsteads (like, 4 or 5 contemporary buildings) seem to be the norm. However, these hamlets seem to be operating within wider regional networks, as they make use of common gravefields, for example. Also, when a particular hamlet is abandoned (due to exhaustion of the nearby field, or vermin, or even the buildings themselves becoming too old, or other less practical reasons), the same settlement location is chosen a few cycles later (a century or so, for example), suggesting these farmsteads operated within a larger rotary system. An elaboration on this settlement system can be found in publications by Stijn Arnoldussen, such as this one. In Denmark, during the Roman Iron Age a shift is visible towards greater nucleation and an increase in settlement size and hierarchy. The archetypical sites here are Hodde, Grøntoft, and primarily Vorbasse. Note that the Romans themselves probably had to do more with aristocratic centers, which are villages/farmsteads (or more accurately, estates) onto themselves; these are not typical for the normal settlement pattern. Tissø is a good example of this.