Visualisation of power-law network topologies

ODL for visualising large network topologies. The main contribution of our algorithm is to simplify the layout problem by separating the nodes in the network into multiple hierarchical layers based on the outdegree of each node. Our algorithm is designed to exploit the underlying structure of power-law topologies, which occur in a wide variety of practical network applications. However, the use of our algorithms is not limited to this class of networks. We have demonstrated that our algorithm can generate useful and aesthetically pleasing layouts for a wide variety of networks, including both regular and power-law topologies. In particular, our algorithm achieved substantial performance improvements over existing layout techniques, including a speed-up factor of up to 260 on real-life Internet routing topologies. I. INTRODUCTION There is a growing need for efficient algorithms to visualise large network topologies in telecommunications. For example, graph layout algorithms can be used in telecommunication network management to help visualise relationships between large numbers of network elements [6], such as the routing relationships between Internet routers [2]. In practice, network managers need to monitor these networks in real-time in order to detect problems due to equipment failures, misconfigurations or unusual traffic patterns. This raises the problem of how to scale-up existing graph layout algorithms in order to visualise large network topologies in real-time. In this paper we present a novel hierachical algorithm, which requires significantly less computation time to layout large networks in comparison to previous approaches [5, 7, 9, 12]. We have implemented this algorithm in a tool that is specifically designed to help network managers analyse large communication networks, such as routing topologies in the Internet. A promising approach to reducing the complexity of the graph layout problem is to use a hierarchical layout algorithm [12]. In a hierarchical algorithm, the given network is first abstracted by clustering or pruning nodes, and the layout of the