Generation models such as Transit-stub [5] and Tiers [8], are centered around reproducing structural properties of the Internet. In particular, Transit-stub has a well-defined hierarchy representing transit and stub autonomous systems in the Internet. Tiers is based on a three-level hierarchy of the Internet as represented by wide-area, metropolitan-area and local-area networks.
Producing synthetic topologies that possess similar structural characteristics to the Internet is important since such properties reflect how the Internet is engineered. On the other hand, achieving hierarchical similarities should not be accomplished at the expense of accuracy with respect to other properties such as degree distributions. There must be a generation model that strikes a good balance between structural properties and degree-related properties. However, it has yet to be developed.
BRITE currently supports generation of two-level hierarchical
topologies. The two-level limit might be overcome by recursively
generating a 
-level topology in 
``phases''. However,
two-level hierarchical topologies are in concordance to the two level
routing hierarchy that has persisted in the Internet since ARPANET
evolved into a network of networks interconnecting multiple autonomous
systems. We plan to extend BRITE to natively support more than two
levels if we find that it would allow for the generation of topologies
that actually reflect real-world scenarios.