The modeling and analysis of large-scale complex systems, such as the Internet, has recently become a hot research topic. We propose a Structure-Based Internet Topology gEnerator (S-BITE) aimed at accurately reproducing the Internet at the Autonomous System (AS) level. The proposed generator exploits a technique that partitions the network topology into two distinct blocks: the Core, which captures the underlying community structure of the Internet, and the Periphery, representing the “tendrils” of the topology. The benefits of this innovative technique are twofold. First, it deals with the high heterogeneity of the Internet by highlighting a small yet well-structured core. This leads to a huge reduction in complexity and shows that the core of the large-scale Internet is not that large, and can further be broken down into a two-layer graph. Second, thanks to the simplifications introduced by the topology layering, it leads us to the definition of a new topology generator, first at the core level and then for the whole Internet.
To the best of our knowledge, S-BITE is the first generator that successfully targets the problem of both matching classical graph metrics, such as the degree distribution, and representing the Internets structure, in the form of maximal cliques. A comparison shows how S-BITE outperforms the reference generators in the literature when looking at both statistical and structural properties of the Internet.