Pregel is a system internally developed by Google, the companion paper of which is very accessible and available for download at http://www.dcs.bbk.ac.uk/~dell/teaching/cc/paper/sigmod10/p135-malewicz.pdf. It represents an efficient, iterative graph-parallel compute model that allows one to implement a large class of graph algorithms. GraphX's implementation of Pregel differs slightly from the preceding paper, but we can't go into any details of this.
In flavor, GraphX's Pregel implementation is very close to aggregateMessages, but has a few key differences. Traits that are shared by both approaches are the send and merge message mechanics. On top of that, with Pregel we can define a so-called vertex program vprog that is executed ...
