Weikuan Yu, associate professor of computer science and software engineering, is increasing the efficiency of computer systems that process big data. From international information technology corporations to the internet search engines we routinely use, Yu’s work has the potential to accelerate the amount of data processed at one time, with one software product already released to Mellanox Technologies for the quick analysis of unstructured data.

With funding from the National Science Foundation, Yu is currently exploring methods to achieve both system and task efficiency on high performance cluster computers. The project focuses on advancing the programming model MapReduce, used by leading corporations such as Yahoo and Facebook to process big data sets with multiple computing jobs simultaneously.

Yu and his research team are creating a set of techniques across multiple layers of software packages by exploring cross-layer cooperation techniques to achieve system efficiency. He is also investigating cross-phase techniques to enhance job fairness and system throughput which enables data to be analyzed at every phase of the software package. To maximize the resource utilization and completion rate of processing multiple tasks, Yu is studying cross-job task co-scheduling techniques that recognize the relationship between jobs to process them in a productive manner. A wide variety of federal and industrial sponsors of Yu’s big data research include the NSF, Lawrence Livermore National Laboratory, Oak Ridge National Laboratory, Mellanox Technologies, Intel and Scitor.