Research On Community Structure And Topology Modeling For IPv6Network Topology

AS the maturation of Internet and rise of the Internet of things, the existing IPv4network is becoming more and more powerless in in resoling the problem of address scale and security and this will undoubtedly accelerate the IPv6network worldwide popularity. In order to make IPv6network provide better service for human society, it’s necessory for us to analysis the topology of IPv6network and design model of IPv6network for research on virus transimssion and security.This thesis take community structure (key concept in complex network theory) as access point and analysis the topological feature of IPv6network and design more reasonable network model that coincide with the real network well.Firstly, based on the authoritative IPv6IP-level network’s data from Dec2008to Feb2011provided by CAIDA, we analysis the basic characterisctics and evolving law of them of IPv6IP-level network with the complex network theory. We reach some conclusion as following:First, the size of IPv6network is growing as the time pass by. Second the distribution of the degree in IPv6IP-level network follows the power-law like many other complex networks. At last the average shortest path in IPv6is larger than that in Internet AS network.Secondly, we compared three common community-detecting algorithms, and then proposed a modified CNM algorithm based on sub-community size. In comparasion with the result of the two algortithm (CNM and CLCNM) in IPv6IP-level network, we reach the conclusion that the performance of CLCNM is better than CNM’s.Thirdly, with the result of community division of IPv6IP-level network topology, this paper analyzed the commmunity features of IPv6IP-level network. From the above analysis, we come to the conclusion that the distribution of sub-community’s scale follows the power-law patterns. And then we define prefix-similarity and get the conclusion that sub-community has close relationship with geographical position. And at last of chapter4, we defined community network to analyze the connection of different regions. And we find that the community network is similar to original network in aspects like degree distribution, coreness and network’s assortativity. At last of this thesis, we proposed a topology model, DCTM (Dynamic Community-detecting Topology Model), based on dynamic community detecting. And we compare the DCTM model, real network and other existing models in degree distribution, network’s assortativity, community structure and etc. And we reach the conclusion that DCTM model performs better in degree distribution and community features than other models. And the conclusion indicates that DCTM model can similate the real network very well.