Rearch On Visual Analysis Of Network Flow Data

With the rapid development of society and science, computer network has become a huge system that covers almost every corner of the world. With the increasing scale of the computer Internet, network communication data, which is generated every time with geometric explosive growth, directly promote the society entering into the era of big data. And network communication data has an irreplaceable role in many areas. However, in the present stage, the utilization of the network communication data is not very ideal with bad real-time and high rate of misunderstanding. While visual analysis technology is increasingly getting mature, its powerful auxiliary analysis capabilities reflected more fully. So, visual analysis of network communication data is a very meaningful for research.Since the effect of visual analysis for network communication data with single method has great limitations, so this paper analyze the data of network communication mainly through three visualization methods, in order to achieve a more comprehensive understanding. The research work of this paper mainly has the following several points:(1) According to the demand of users for analyzing the communication accessibility relations between the nodes in a network, proposed a kind of parallel coordinates visualization method with clustering based on communication strength. Show the interested node, source nodes collection, the target nodes collection and distribution of them, as well as the reachability relation between them, in a unified view, through the definition of reachability matrix, communication intensity matrix. Cluster the reachability curves with fuzzy method, according to the three level strategies.(2) According to the demand for analyzing the special communication behaviors, summed up the inherent communication patterns of them. Designed the visualization scheme based on parallel coordinates plot. Using the different patterns of these communication behaviors, help users to find out the special behaviors from complex flows.(3) Conventional network topology can only show the relationship between nodes, but can hardly support the following analysis of business structure. This paper proposed a layout algorithm based on flow-guiding force, showing not only the connection between nodes, but also the attributes of low and class of nodes.(4) According the demand for analyzing the carrying relationship of flows, proposed a visualization method scheme based on the hierarchical layout. Partially adjusted the existing layout according the corresponding relations and differences between layouts, and showed the multi-level structure and the mapping relations in a uniformed view, helping the users to visually analyze the carrying relationship of flows.