Research And Implementation Of Network Topology Three-dimensional Visualization System

The rapid development of the internet makes the sharp increase in the number of network nodes and the increasing complexity of connection between those network nodes, which makes it difficult for network managers to understand and analyze the network. Network visualization uses the knowledge of computer graphics to change abstract data information into intuitive graphics for analysis, helps network managers understand the network more accurately and effectively, so the research is very important.Node layout algorithm is the core of the network visualization. The shortage of existing layout methods is mainly in three aspects: One is that some layout methods like the hierarchical layout algorithm can only handle simple tree topologies, when dealing with complex mesh topologies, it is difficult to get a clear visual graphic; the second is that some algorithms like the force-directed algorithm have a high time complexity, it can not process networks with massive number of nodes; the third is that most of the layout methods can not reflect the internal structure of the network. In order to solve the problems above, this paper uses an algorithm that place node in hyperbolic space, and improve the algorithm with community division so that the layout result can reflect the true geographic information of nodes. The main work of this paper includes:(1) Research and implementation of node layout algorithm in hyperbolic space. in hyperbolic geometry the space of hyperbolic grows exponentially, this property just meets the fact that the number of node increases in network topology with the increase of network-level, the high layout efficiency combined with the property of hyperbolic space make the algorithm more capable of handling networks with massive number of nodes. This paper implement two-dimensional and three-dimensional node layout algorithm in hyperbolic space, and use different data to test it, the results show that the algorithm can reflect network topology well.(2) Proposed a location-based constrained three-dimensional hyperbolic space layout method. This method is improved upon three-dimensional layout algorithm in hyperbolic space, The original three-dimensional hyperbolic space layout algorithm can only handle tree topology, this paper combining it with label spread communitydetecting algorithm, we can get a tree topology which can represent the structure of the whole mesh topology according to the community detecting result, and use this tree topology to layout the network. Test result show that the new method which combined with community detecting algorithm can not only handle complex mesh network topology efficiently, but also make visual graphic good enough to reflect the real structure information of the network topology.Based on the above studies, this paper presents a three dimensional network topology visualization system based on three dimensional hyperbolic space layout algorithm. The system not only integrates three-dimensional hyperbolic layout algorithm and its improved algorithm as the core of data processing module and visualization graphics rendering module, but also includes visualization graphics human-computer interaction like zoom in and out, translation and rotation. We tested the functions of the system, the result shows that the network topology three-dimensional visualization system can well complete visualization related tasks.