关键词: 可视化, 4over6, 流量特征, 派生数据

Abstract: Flow visualization is of great help for network researchers and supervisors to detect network anomalies, understand the characteristics of network traffic and find time trend, etc. Existing flow visualization technologies have failed to fully meet the growing needs of traffic visualization. Aiming at time, dimension and structure characteristics of flow, this paper adopts spiral visualization technology, dimension classification method and space filling technology to meet the need of different characteristics of flow. By visualizing the data from 4over6 project and compared with the methods currently used in the project, the results show that spiral visualization can excavate the time characteristics of traffic data more effectively than line mapping; dimension classification is more effective for special dimensions visualization than parallel coordinates and space filling can express node information with no loss of structure characteristics which node-link mapping can’t achieve.

Key words: visualization, 4over6, traffic characteristics, derived data