PennAnalytics: Network Visualization and Analytics

Abstract

Society is becoming increasingly digitized and reliant on computer networks for everything from business to leisure. Due to this reliance, it has become increasingly important for computer networks to be reachable around the clock. Meanwhile, computer networks are growing in size and complexity. With the importance, ubiquity, and growth of computer networks, issues such as congestion are becoming more prevalent and increasingly challenging to address for network administrators who are responsible for managing computer networks. As a consequence, the need for easy-to-use tools to analyze computer networks is more urgent than ever. Many current network analytics tools present data in a way that is hard for humans to understand and use effectively. PennAnalytics is a tool designed to alleviate this situation and help IT professionals better understand the condition of a computer network. It uses Simple Network Management Protocol (SNMP) to visualize and analyze computer network data in real time and display information in a user-friendly way through a web browser. PennAnalytics differentiates itself in three dimensions: realtime analysis, cross-platform compatibility through web browsers, and a focus on user experience. Instead of manually building out a network to visualize connections between nodes, PennAnalytics automatically builds the network graph and displays traffic over time. It was designed as a tool to allow administrators to quickly understand problems at a macro level, as well as easily drill down and display potential problems at the micro level to expedite and enhance the manual network diagnostics process. In order to do this, we built the entire stack that supports a tool that visualizes the network and network metrics in a way that provides clarity on network problems such as traffic spikes and network congestion.

View Slides

4 Figures and Tables

Cite this paper

@inproceedings{Wingo2014PennAnalyticsNV, title={PennAnalytics: Network Visualization and Analytics}, author={Patrick Wingo and Aubrey Chase and Bill He and R. Sasson}, year={2014} }