Abstract:
In this talk, I will present an overview of PRIME, an ongoing research project at Colorado School of Mines sponsored by an NSF CAREER Award. The design goal of PRIME is to provide a self-sustained large-scale virtual network environment for researchers to prototype, evaluate, and analyze distributed applications and network services. Built on the previous success of SSFNet, PRIME achieves high-performance real-time network simulation by taking advantage of both parallel simulation and multi-resolution network modeling techniques. I will focus on two problems at hand and present our solutions. One is the integration of an efficient fluid TCP model based on ordinary differential equations with the packet-level simulation using discrete events. Our scheme allows simulation to dynamically change the composition of traffic flows to balance modeling accuracy and performance. The other area is the design of an open and scalable emulation infrastructure that can easily incorporate real-world distributed applications and network services to interact with the simulated network. A prototype of the emulation infrastructure has been implemented based on Virtual Private Network (VPN) bridging traffic between the physical entities and the virtual network. I will also provide a summary of other current research areas and future directions.

Biography:
Jason Liu is an assistant professor of Computer Science at the Colorado School of Mines. Prior to that he was a post-doc at the Coordinated Sciences Laboratory at the University of Illinois, Urbana-Champaign. He received a B.A.degree in Computer Science from Beijing University of Technology in China in 1993, an M.S. in Computer Science from College of William and Mary in 2000, and a Ph.D. in Computer Science from Dartmouth College in 2003. His research interests include parallel discrete-event simulation, high-performance modeling and simulation of communication networks and computer systems. His current research focuses on applying real-time computation techniques for adaptive network simulation models, designing and building scalable emulation infrastructure for large-scale network simulations.