Abstract:
Virtualization technologies are playing an increasingly significant role in today's computing systems, spanning all layers of the hardware/software stack, from virtual storage, virtual network, to virtual machine and virtual service. Common among these technologies are their unique capability of enabling important resource abstractions needed for efficient, secure, and reliable computing. This talk will use two specific examples of virtualization in my research to demonstrate the power of such an approach.

The first example considers virtual machine based datacenter resource management. It leverages virtual machines to provide flexible resource sharing and employs intelligent resource controllers to satisfy application service quality requirements. One particular interesting application supported by this virtual resource management system is the parallel modeling for brain-machine interfaces. The second example is on the virtualization of distributed file systems for grid data management. It enables on-demand cross-domain data access for unmodified applications with tailored enhancements on performance, security, and fault-tolerance. This approach has supported the grid computing of scientific applications from many domains, including the biomedical spectroscopy analysis for early cancer detection.

The exciting results obtained from the above research also motivate us to build tomorrow's large-scale virtualized systems for high-performance and high-throughput computing. This talk will be concluded with the discussions about the research challenges and my future work towards this goal.

Biography:
Ming Zhao joined the School of Computing and Information Sciences at Florida International University as an Assistant Professor in fall 2008. His research interests include distributed computing, virtualization, high-performance systems, and autonomic computing. He received his B.S. and M.S. in Automation from Tsinghua University, and his Ph.D. in Electrical and Computer Engineering from the University of Florida.