SC21 Proceedings

The International Conference for High Performance Computing, Networking, Storage, and Analysis

Yin and Yang: Balancing Cloud Computing and HTC Workloads


Student: Zhuangwei Kang (Vanderbilt University), Zhuo Zhen (University of Chicago), Kate Keahey (Argonne National Laboratory (ANL))
Supervisor: Zhuo Zhen (University of Chicago)

Abstract: With the help of proper preemption policies and proactive resource schedulers, combining academic cloud and High Throughput Computing(HTC) systems through preemptible instances would help increase the utilization rate of the clouds and reduce the energy cost at the same time. We propose a data-driven simulator CHISim, then with which we evaluate a comprehensive set of resource scheduling strategies (4 preemption policies × 3 cloud user requests forecasting models). By exploring these strategies, this paper seeks to answer the following question: (1) Can we fill cloud resources lease gaps by deploying HTC jobs? (2) What is the most efficient strategy to do this? (3) How can we minimize the cost of doing this by adopting strategies that reduce HTC reruns?

ACM-SRC Semi-Finalist: no

Poster: PDF
Poster Summary: PDF


Back to Poster Archive Listing