Empirical Study of the IPv4 and IPv6 Networks at the Autonomous System Level.
Jiaqi Nie,Jiangtao Hao,Songtao Peng,Xincheng Shu,Zhongyuan Ruan,Qi Xuan
DOI: https://doi.org/10.1109/DSC59305.2023.00087
2023-01-01
Abstract:In this paper, based on the recent data (up to July, 2022), we empirically investigate the evolution of the IPv4 and IPv6 Internet topologies at the autonomous system (AS) level. We find that both networks are growing with time, while the IPv4 network is much greater in scale (including the number of nodes and links), but smaller in link density (average degree). The cumulative degree distribution of each network exhibits a double power-law form, which is different from the early observations. Moreover, we show that the innermost cores (obtained by the k-core decomposition algorithm) in two networks are largely overlapped. Finally, by inspecting the clustering coefficient and the degree correlations, we demonstrate that the two networks are hierarchically organized. Another traditional indicator, i.e., the normalized rich-club coefficient proposed by Colizza et al., fails to capture the hierarchy property of the IPv4 and IPv6 networks. Nevertheless, if we generalize the concept of “rich club” by evaluating the richness of a node in terms of “coreness” (not in terms of “degree”), we still observe such phenomenon at some times.