UNC Network Vis

Overview

force_net_viz_smThis image is a graph-based network visualization consisting of traffic from UNC department subnets to external machine connections.

The graph is placed in three dimensions using a gravity model where each edge provides an attractive force and each node provides a repelling force. This can be used to variably weight nodes in the graph in order to express particular properties of the graph.

The size of the lines represents the ammount of traffic for this particular time interval, while the color represents the type of the connection: web, file transfer, and so forth. The round, green objects represent the external machines and the squares represent the internal subnet connections, which are randomly colored.

Each node in the graph is labeled with the IP address as well as the internal subject if applicable. To the lower left we provide a histogram detailing the amount of traffic through a selected subnet. To the right we provide a table showing the color by which we differentiate the UNC subnets.

renci_netviz_01The second image represents the UNC network traffic via a georeferenced visualization where each UNC subnet is assigned a geographic location via a green sphere. Six-inch-resolution LIDAR data was used for the terrain, while high-resolution orthophotos were used as texture information to represent the UNC campus.

When a connection exists between each subnet, an arc is drawn from one ground location to the connecting location. Each connection is then colored based on based on type (web, file transfer, and so forth). When more than one connection exists of differing colors, the connections are bundled by packing them along the edge of a circle in order to keep the image from becoming cluttered.

Connections from on-campus to off-campus machines are routed from the geographic location to a fixed location in space above the terrain, which represents all outside traffic. Connections are once again colored based on the type of connection.

Project Team

  • Jason Coposky
  • Bryan Byerly