Wireless Simulation (A6)
Project PI: Gustavo de Veciana
PROJECT DESCRIPTION
A growing core of telecommunications research faculty at UT-Austin are addressing problems ranging from physical layer transmission issues to network layer protocols and traffic management. This work increasingly relies on simulation and computer based engineering tools to support both teaching and research. Significant challenges arise from the introduction of new communications technologies, increased algorithmic complexities, and the growing interdependence between computing and communication, which make the integration of high performance, networked workstations indispensable.
This project comprises collaborative research in three areas: 1) modeling and development of smart antenna array systems for wireless communications; 2) development of a Statistical-Match Queuing (SMAQ) tool for measurement-based traffic modeling and evaluation of network management strategies; and 3) development of a simulation platform for evaluating the macroscopic behavior of large-scale hierarchical routing algorithms. As discussed below, in order to tackle increasingly complex channel characteristics and telecommunication systems, our work relies heavily on computational resources and tools, and it is our intent in migrating towards a networked Intel/WindowsNT computing environment to effectively serve the needs of the increasing number of students, faculty, and research staff in the Telecommunications area.
TECHNICAL CHALLENGES
In order to properly assess the performance of current communication systems, our research makes use of techniques based on processing huge amounts of real data for traffic and wireless channel modeling, e.g., a week of processing for a days worth (Gbits) of experimental data. Assessing system performance based on real traffic traces is a slow and inflexible process, so our researchers have developed tools for automatically generating Markovian models which match real traffic traces. The huge processing costs of obtaining such models and then computing system performance, payoff handsomely by allowing engineers to explore synthetic aggregation, multicast, and design/control options for networks subject to various types of traffic mixes. Such tools enable engineers, to effectively design network systems with huge capacities where simulation is no longer an option.
IMPACT
Our researchers, have close ties with leading telecom industry players, and are broadly supported by government agencies, including, Texas Instruments, General Electric, Motorola, Southwestern Bell, Nortel, DARPA and NSF. The research, tools under development, and increasing telecom student body, are geared to meet many of the critical problems in field today, such as the limited capacity of wireless systems; assessing the performance of hybrid switching architectures for Gigabit IP routers; and the exploring the dynamics of future hierarchical (PNNI) routing standards for large-scale ATM networks.
EQUIPMENT
We propose the following equipment request aimed providing networked high-end compute servers to our researchers: 2 Quad Pentium Pro Servers; 2 Dual Pentium Pro Servers; 8 Dual Processor Graphics Workstations; 3 Pentium Pro Workstations and a Switching Hub to be acquired over a two year period.
RESOURCES
This equipment, our current space allocation, and support from our funded research will enable us to develop a networked computing environment shared by 5 faculty and over 30 graduate students they supervise.
BENEFITS TO INTEL
Future telecommunications engineers will rely heavily on high-end workstations for the development and prototyping of algorithms, software, and hardware for communications systems. By supporting this work, Intel will benefit, by putting the power and flexibility of the Intel based equipment to the test in complex models and development of new engineering tools using the Intel/WindowsNT platform, giving further exposure to the telecommunications academic/industry community.