Title :
Bipartite Internet topology at the subnet-level
Author :
Akgun, Mehmet Burak ; Gunes, Mehmet Hadi
Author_Institution :
Dept. of Comput. Sci. & Eng., Univ. of Nevada, Reno, NV, USA
fDate :
April 29 2013-May 1 2013
Abstract :
Internet topology modeling involves capturing crucial characteristics of the Internet in producing synthetic network graphs. Selection of vital metrics is limited by our understanding of the Internet topology, which relies on the state of the art measurement studies. Recent measurement studies indicate that underlying subnetworks, multi-access links providing one-hop connectivity to multiple nodes, are an important factor shaping the topology of the Internet. Current network models utilize point-to-point edges that can connect exactly two vertices of the topology. Decomposition of the underlying multi-access links into pairwise point-to-point edges results in cliques and is an oversimplification of the analyzed networks. Accurate modeling of multi-access links require special type of edges, i.e. hyper-edges, that can connect multiple vertices in a hyper-graph. Hyper-graphs are best illustrated as bipartite graphs where hyper-edges connect two types of vertices, i.e., router interfaces and subnets. In this paper, we introduce a bipartite model of the Internet topology and discuss representative synthetic network generation.
Keywords :
Internet; graph theory; multi-access systems; network theory (graphs); telecommunication links; telecommunication network routing; telecommunication network topology; Internet characteristics; bipartite Internet topology modeling; cliques; hyper-edges; multiaccess link decomposition; multiaccess link modeling; one-hop connectivity; pairwise point-to-point edges; point-to-point edges; router interfaces; subnet-level; synthetic network generation; synthetic network graphs; topology vertices; Artificial neural networks; Generators; IP networks; Internet topology; Network topology; Topology; 2-mode; Hyper-graph; Internet topology;
Conference_Titel :
Network Science Workshop (NSW), 2013 IEEE 2nd
Conference_Location :
West Point, NY
Print_ISBN :
978-1-4799-0436-5
DOI :
10.1109/NSW.2013.6609200