This project provides visualizations on an as needed basis in support of the larger Foundations of Measurement Science for Information Systems project. It aims to develop and evaluate a coherent set of methods to understand behavior in complex information systems, such as the Internet, computational grids and computing clouds.
Our goal is to help network researchers with insight to the functioning of the networking systems being investigated, primarily via a simulation. Visualizations of these complex simulations has provided, at a minimum a secondary independent method to ensure the proper execution of the simulations, and more fruitfully as an aid to insight about the simulation.
As an extension of earlier work we produced an animation of a network based on the 2001 configuration of the Autonomous System (AS) portion of the Internet, the team uses for study. The graph consists of approximately 11,000 nodes and 47,000 edges. Using an off-the-shelf open source tool called "gephi" we produced a number of visualizations that have helped give some structure to the network, which was completely unknown before seeing it. Data illustrated using gephi shown in the following figure was animated using alternate visualization parameters, see frames that follow. Colors correspond to data clusters using the "modularity" network metric.
Internet Autonomous System Graph Circa 2001 in gephi tool
Frames from animation of Internet Autonomous System Graph Circa 2001
Topological and bar graph visualization of cloud simulation data. Available online here.
Continued work by Mills and Dabrowski has produced a network simulation that examines the role of realistic parameter changes as it affects the performance of a network. As part of work as a guest researcher from CSIRO in Australia, Phil Gough produced a D3-based interactive visualization tool for network simulation data produced by this simulation. This tool allows network researchers to look in a detailed manner at the results of simulations and to easily identify conditions such as network congestion.
Network simulation visualization tool.
Topology on the left side of the display illustrates queue lengths and congestion state. Column type displays on the right side allows a user to select parameter input to simulation and displays congestion for each type of node group. The graph on bottom displays congestion and lets user step through the time of simulation.
A live demonstration of this tool is available here. Currently two datasets are available one with 250 time steps, the other with 3000 time steps; simply use the desired number in the URL.
Network simulation visualization tool details.
As illustrated in figure 2 there are five groups or types of nodes (indicated by color of interior of circle); Node congestion or occupancy is indicated by color of border of circle; Two small circles inside of node indicates queue length of up (top) or down (bottom) queue. In the future we expect to produce some additional animations based on this visualization and collaborate on an as needed basis with network researchers. This work was also part of Gough's research on communicating scientific results to non-scientists.
Return to Visualization
Return to Visualization