I've been reading on distributed systems for processing on large graphs. The most prominent examples include Pregel (developed by Google) and Apache Giraph. Most of these systems argue their existence that they are for "big data" processing, i.e., they can handle graphs that are too big to fit into a single machine.
While I'm aware that there are (a few) real world examples of graphs that are HUGE, for example, the Facebook graph (billion of nodes, trillion of edges), or the web graph, it seems to me that these are outliers and most people will rarely ever have to deal with such data sets.
Are there examples of graphs that frequently occur in the real world (i.e. either created by humans or occurring in nature) that are so huge that they cannot be handled by a single (commodity-type) machine?