EECS Faculty Candidate Seminar- Dr. Tevfik Kosar
Green Data Flow: Energy-efficient Data Transfer Scheduling and Optimization
Dr. Tevfik Kosar, University at Buffalo (SUNY)
2:00pm, Thursday, February 28
Min Kao room 435
With the emergence of complex scientific applications, social media, video over IP, and more recently the trend for Internet of Things (IoT), the global data movement requirements have already exceeded the exabyte scale. It is estimated that, in 2019, more IP traffic will traverse global networks than all prior "Internet years" combined. To support this massive amount of data movement across the globe, the networking infrastructure as well as the source and destination nodes at the edges consume immense amount of electric power, with an estimated cost measured in billions of dollars. Although considerable amount of research has been done on power management techniques for the networking infrastructure, there has not been much prior work focusing on power-aware data transfer algorithms for minimizing the power consumed at the end-systems. In this work, we analyze various factors that affect the power consumption in end-to-end (cloud-to-edge and edge-to-edge) data transfers, such as the level of parallelism, concurrency, pipelining, CPU frequency level, and task scheduling at the end-systems. We introduce novel data transfer algorithms which aim to achieve high data transfer throughput while keeping the energy consumption during the transfers at the minimal levels. Our novel solution combines historical data analysis with real-time online learning and achieves significant amount of energy savings (60-80%) at the end-systems during data transfers without sacrificing from the performance. In many cases, we can achieve performance increase and save energy at the same time.
Tevfik Kosar is an Associate Professor in the Department of Computer Science and Engineering at the State University of New York at Buffalo (UB). He has received his Ph.D. in Computer Science from the University of Wisconsin-Madison in 2005. His main research interests include data-intensive distributed computing, big-data analytics, datacenter networking and I/O optimization, performance and energy efficiency in IoT and edge computing systems. He is the primary designer and developer of the Stork distributed data scheduling system, and the lead investigator of the PetaShare distributed storage network. Some of the awards received by Dr. Kosar include NSF CAREER Award, IBM Research Award, Google Research Award, LSU Rainmaker Award, LSU Flagship Faculty Award, Baton Rouge Business Report’s Top 40 Under 40 Award, 1012 Corridor’s Young Scientist Award, UB Senior Faculty Teaching Award, and UB Senior Faculty Research Award. Dr. Kosar has published more than 100 scholarly articles in top journals and conferences in his area. He consistently serves on the PC of related conferences such as IEEE/ACM SC, ACM HPDC, IEEE CLUSTER, IEEE ICDCS, IEEE CLOUD and IEEE BigData. He is currently serving in the Editorial Boards of the IEEE Transactions on Parallel and Distributed Systems, and the Cluster Computing journal.
Thursday, February 28 at 2:00pm to 3:00pm
Min H. Kao Electrical Engineering and Computer Science, 435
1520 Middle Drive, Knoxville, TN 37996