Spotlight: Lei Lan

Lei Lan is a former Master's student of the University of Windsor in Computer Science. During her studies in Windsor, Lei Lan specialized in systems software for High Performance Computer under supervision of Dr. Angela Sodan. Lei Lan was supported by an OGS scholarship and Windsor tuition scholarships as well as NSERC grant money.

Lei Lan did her research work on a novel job scheduling approach for parallel applications that improves the utilization of the machine significantly and subsequently provides much better response times. The approach exploits simultaneous execution of multiple jobs on shared resources, whereas the conventional approach is to allocate resources in a dedicated manner. The scheduling approach developed by Lei Lan partially reorders the queues to find jobs that are good candidates for simultaneous execution by having different characteristics in I/O, computation, and communication, and keeps an eye on the effect which the reordering the queue has on the response times. The suitability and improvement potential for real applications is shown by tests with the NAS benchmarks which are a standard benchmark suite to test the usefulness of system optimization.

Overall, the approach obtained an improvement in response times by up to 50% which means a significant improvement in the utilization of the expensive HPC resources and a significant improvement in user satisfaction.

"I really enjoyed this experience and am happy that I had the opportunity to work on such an interesting topic with Dr. Sodan. I did not even think of relating my name to an IEEE paper before." says Lei Lan. Her Master thesis work was published in IEEE Transactions of Parallel and Distributed Computing which is the top journal of the field.