Community Structure Detecting Of Multiple Granularity And Visualization Based On Internet Network Topology

The community studies of complex network have relatively matured, but the study for community based on characteristics of the Internet topology so far is relatively lacked, and no literature studies the Internet’s unique structural features. Community structure of the system is closely related to the functions. a deep understanding of the characteristics of the Internet community structure can help people find the Internet’s new functional unit, understand the complex network characteristics of the structure, simplify network structure, look for a new strategy which combats Internet viruses to spread and control network.As to the unique structural features of the Internet, such as tree feature of the edge of the Internet network, the Internet chain feature (network、BGP external route), the clustering feature of the high core nodes, the central node feature. This paper in the course of the study of the Internet community compares with the traditional complex network of community detecting algorithms, uses the dividing and getting together methods and designs the level folding contraction method、the chain-detection method、node clustering feature of the high core detection method、the central node method. These methods can be used to roughly find the Internet smaller functional organizational structure, for example, the level folding contraction method can detect the edge of the tree structure feature organization of the Internet network, chain detection method can detect the structural method organization of the Internet submarine cable, network, chain characteristics organization like external route of BGP, high core node clustering detection method can detect organization with the high core nodes clustering with each other feature in some areas, the Center node method can detect the feature organization of the Internet center node. Through these detection method, most of the network nodes will be classified to the appropriate organizational structure, but some small part of the nodes may be not in above several features organization. The above detection methods are unable to classify these nodes to the appropriate organizational structure, so they are classified as a separate organizational structure (to facilitate in the follow-up processing). After processing with the front steps, a whole organizational structure of the entire network topology is revealed, the next step is to merge these different sizes of the network organizational structure to become one community with high cohesion and low coupling characteristics, which is named "together". But to make the merged network community modularity to achieve optimal is an NP-hard, the time complexity and space complexity was relatively large. As to the optimization process of community module for many smaller organizational structure of the network community, small communities merge method designed by this article uses a greedy algorithm to find the local optima of modularity, and eventually a network topology with the characteristics of the typical community structure division surfaces. In this paper, the community detect algorithm is made up of a group of multi-granularity community detect method, known as AY algorithm. This algorithm in the IPV6network community is superior to fast unfolding of communities in large networks(referred FUOCILN)algorithm, which is known as beyond many traditional societies detecting algorithm. Nowadays, this algorithm is further optimized and promising.For visualization of the Internet community, some of the traditional visualization algorithms only can do the whole level visualization of network topology, visualization feature of the community structure is not clear. For this problem, this article builds a new graph layout on the foundation of the traditional graph layout; this layout is divided into two big steps:the layout of the entire network community, the layout of the internal nodes in each community. The first step consists of two levels:firstly, according to the physical analogy, it simulates the physical system environment, and automatically layouts community node; secondly, after the first step, it use dynamic interactive layout mode. If the first step has some unreasonable parts, you can manually adjust the location of the Associations node. The complexity of large-scale data and the problem of outstanding association characteristic must sacrifice the canvas resources of the internal layout. For this, the visualization algorithm uses the ray layout algorithm to display the topology of the internal nodes of a particular community and open a new canvas resource, In view of the characteristics of the visualization algorithms, it is called visual Community(referred VC). Visual Community can reasonably and clearly visualize the characteristics of the Internet community structure.