BRITE separates the placement of the nodes from the process of growing the topology and interconnecting the nodes. By placing a node we mean just selecting a location in the plane for it and creating and initializing the data structures for the node in the graph. This phase does not mean that the nodes already belong to the topology because the specific joining time of a node to the topology will depend on the growth model employed.
The class RouterModel provides a method called PlaceNodes that places the nodes on the plane in one of two ways: randomly or heavy tailed. The motivation behind providing heavy-tailed distributions is explained in Appendix A. When node placement is random, each node is placed in a randomly selected location of the plane. When the placement is heavy-tailed, BRITE divides the plane into squares (the size of the plane and the size of the squares is controlled by parameters passed to BRITE. See Section 3.8). Each of these squares is assigned a number of nodes drawn from a heavy-tailed distribution. Once that value is assigned, then that many nodes are placed randomly in the square. Again, this placement mechanism can be modified or overridden by particular models.
Figure 4 shows the difference between random and heavy-tailed node placement. The clustering provided by heavy-tailed placement can be used for specific generation models [16].
