|
MS Group / Home
Modeling and Simulation for Performance and Tradeoff Analysis in High-Performance Computing and
Networking
The modeling and simulation (MS) group has a long and successful history on research in modeling and
simulation for the study of performance and related issues in high-performance networking and computing. The
group has developed a number of models sets, tools, and publications that provide mechanisms and insight for
high-fidelity modeling and fast simulation of networks, architectures, and systems for key HPC
and embedded applications.
For DoD, the group currently focuses on development of simulation models, tools, case studies, and experiments
for the design and tradeoff analysis of cluster and grid networks, architectures, and systems. Although clusters
and grids are becoming a dominant approach for advanced computation, the design space of such system
architectures is becoming exceedingly complex, in terms of the wide array of design strategies for node
architectures, interconnect architectures, interfaces, services, algorithm decompositions, etc. in light of
emerging technologies. This complexity makes unfeasible the simple use of analytical projections, and the costs
involved make testbed experimentation alone an impractical option. Thus, new simulation studies and tools are
needed to better explore the design space and provide for the selection of the most effective and efficient
configurations. This group is investigating ideas and tools to achieve these objectives for a variety of
application.
For Honeywell, the group currently focuses on virtual prototyping of advanced space system architectures based
on RapidIO. This work includes an investigation through computer-based simulation into the optimal means by
which to develop advanced space system architectures based on RapidIO. In addition to the vast design space
regarding node and board architectures in such an embedded processing system, there are a countless number of
options in the RapidIO design space such as link speed, link width, and level of flow control. Also, a clear
path to flight for commercial RapidIO components must be determined. This project is using simulation in order
to gain a thorough understanding of the tradeoffs associated with these design options supporting several
space-based radar algorithms on a space-based cluster of embedded processing nodes.
In addition to work directly supporting requirements for DoD and Honeywell, this group also collaborates with the
HPC group in the lab and their work on scalable,
robust on-line monitoring and management for large-scale networks and network-based systems. These two groups
support efforts in resource monitoring and network simulation for the iVDGL
group at UF to help address their critical requirements from high-speed computer networks and systems.
Principal Investigator: Dr. Alan D.
George
Group Members
Eric Grobelny, PhD student, group leader
David Bueno, PhD student
Chris Conger, MS student
Adam Leko, MS student
Casey Reardon, PhD student, UF Presidential
Fellow
Ian Troxel, PhD student
Related Links
Progress Report on MS research in HCS Lab (10/15/04) -- PPT
Annual Progress Report on MS research in HCS Lab (05/12/04) -- PPT
Progress Report on MS research in HCS Lab (02/05/04) -- PPT
Progress Report on MS research in HCS Lab (11/05/03) -- PPT
Progress Report on MS research in HCS Lab (07/30/03) -- PPT
Eric's site for
MS group materials
Honeywell project
site (password protected)
Optical
Network Group project site
Acknowledgements
We gratefully acknowledge the following vendor for their support of this
project:
MLDesign Technologies for
their donation of software tools.
|
 |
![[the lab]](/img/b_lab.gif)
![[projects]](/img/b_projects.gif)
![[people]](/img/b_people.gif)
