Accelerating Big Data Applications using High Performance Computing Technologies

Wednesday, March 9, 2016 - 10:10
TH 331
Michelle Zhu
This talk highlights some research projects of Dr. Zhu in the area of high performance computing and big data applications. Big data is no longer a buzzword, but a fact of modern business, healthcare, and science life. The cutting-edge high performance computing and networking technologies are required to manage and analyze large volume of data, which may be located in various geographical distributed sites. Efficient resource management and data analytics strategies can help discover knowledge and patterns from this huge information sea across a wide area network. How to utilize dedicated channel reservation on high performance network for data transfer, Remote Direct Memory Access (RDMA) enabled high-performance interconnects, as well as high-speed storage systems to boost the networking and I/O components of the middleware should also be considered. One of her major projects is remote visualization and steering for scientific workflow applications across the wide area network. This system enables scientists to conveniently assemble, execute, visualize, control, and steer computing workflows in distributed environments via a unified web-based user interface. A class of efficient workflow mapping schemes is incorporated to achieve optimal end-to-end performance based on rigorous performance modeling and algorithm design. Furthermore, as streaming applications on Apache Storm or Spark are gaining prominence to process real-time continuously generated business or scientific data, she studies the resource management strategy to achieve the maximal processing rate. Her recent research interest is in green cloud computing. Due to the increasing deployment of data centers around the globe, the electricity cost on the computing, communication and cooling system have skyrocketed. In order to maintain a sustainable Cloud computing paradigm in the future, her group design and develop energy-aware scheduling approaches to minimize energy consumption while still satisfying certain Quality of Service (QoS) such as response time specified in Service Level Agreement (SLA).

Dr. Michelle Zhu finished her dissertation in Computer Science and Mathematics Division at Oak Ridge National Laboratory (ORNL) and received her Ph.D. degree in Computer Science from Louisiana State University in 2005. She joined the Computer Science Department of Southern Illinois University Carbondale as an assistant professor in Jan. 2006 and was tenured and promoted to associated professor in 2011. She currently serves as the undergraduate program director and ABET assessment committee chair for the department. Her research interests include High Performance Computing; Cloud Computing and Big Data applications. She has published about 100 research articles in prestigious journals and conference proceedings including IEEE Transactions on Computers (TC), IEEE Transaction on Parallel and Distributed Systems (TPDS), Journal of Parallel and Distributed Computing (JPDC), IEEE Transactions on Network and Service Management (TNSM) and ACM Transactions on Sensor Networks (TSN). Her research and education projects have been sponsored by Department of Energy (DOE), ORNL, National Science Foundation (NSF), Illinois State Board of Education and NVidia.