our research


Nanocubes (YouTube)
Nanocubes are an efficient in-memory data structure for visualizing billions of geospatial and temporal data points at interactive rates.

MINGLE - a Multilevel Edge Bundling Algorithm (mpeg - 25 MB)
Edge bundling is a technique that can help in reduce clutters of large graph visualization. The MINGLE algorithm can scale up to millions of edges.

LiveRAC (mpeg - 30 MB)
LiveRAC is a visualization system that supports the analysis of large collections of system management timeseries data consisting of hundreds of parameters across thousands of network devices. It provides high information density using a reorderable matrix of charts, with semantic zooming adapting each chart's visual representation to the available space. LiveRAC allows side-by-side visual comparison of arbitrary groupings of devices and parameters at multiple levels of detail. The video illustrates its capabilities.

Path Router (mpeg - 4.3 MB)
Illustrates a technique for routing a smooth curve between two points while avoiding intervening objects. The technique and its related aesthetics were designed for drawing edges of graphs.

Hardware-accelerated View-dependent Map Simplication (Quicktime - big! 200 MB)
Demonstrates a novel hardware accelerated map simplification that improves rendering performance and reduces clutter in interactive map viewers.

Kinetic Depth Contours (Quicktime)
Demonstrates the power of graphics hardware for computing the depth contours of a set of points in the plane. The video is a real time capture of the computation. Also see the paper "Hardware-Assisted Computation of Depth Contours" by Krishnan et al.

Bounding Box of Moving Points (Quicktime)
Demonstrates the use of duality in hardware to compute the minimum enclosing bounding box of moving points in three dimensions. This video is based on work done by Shankar Krishnan and Suresh Venkatasubramanian with Prof. Pankaj Agarwal and Nabil Mustafa at Duke.

Toplogical Fisheye Viewer demo (Quicktime)
A rough cut demonstrating the Topview user interface for a browser for large graphs. We illustrate it on the (inferred) peer-to-peer core of an AS graph hierarchy (from work by Agarwal, Subramanian, Rexford and Katz), and on a 148,000 node router graph provided by Bill Cheswick formerly at Lumeta Corp..

Case Studies

AT&T Visualizer (Vizgems)

The AT&T Visualizer is a service offered to enterprise customers that provides visual monitoring and interactive querying of their managed services (Internet hosting, WiFi, managed routers) in our Internet Data Centers within the AT&T Business Direct Portal. This service runs on a data acquisiton, integration and visualization platform created by Eleftherios Koutsofios in our group. It scales to hundreds of millions of events per day in realtime without relying on sampling or aggregation.


In 2008-09, the LiveRAC system for visually exploring time series data sets was generalized to work with several different relational database systems, including the Daytona system that stores CPU and memory statistics for monitoring MPLS routers in AT&T's IP backbone network. This was joint work with Jennifer Yates in the Internet and Network Systems Lab.