Large-Scale Network Simulation

Wednesday, November 5, 2003 - 3:15pm - 3:50pm
Keller 3-180
Richard Fujimoto (Georgia Institute of Technology)
Title: Large-Scale Network Simulation Abstract (five lines to a half page): Parallel and distributed simulation tools are emerging that offer the ability to perform detailed, packet-level simulations of large-scale networks. This capability offers enormous new opportunities for researchers to perform simulation experiments on networks of a scale that could not be completed previously. At the same time, it also creates challenges to the research community to define scenarios and configurations that are realistic relative to current and future Internet configurations. It creates challenges to tool builders to develop verified and validated simulators that are easy to use and which execute efficiently on parallel and distributed computers over a wide range of network configurations and scenarios. This presentation will quantitatively characterize the state-of-the-art in large-scale network simulation. An approach to realizing scalable network simulations that leverages existing sequential simulation models and software will be described. A recent performance study is described concerning large-scale network simulation on a variety of platforms ranging from workstations to cluster computers to supercomputers.

This research represents joint work of the speaker with Drs. Mostafa Ammar, Kalyan Perumalla, George Riley and several graduate students at Georgia Tech.