Cost-aware Parallel Scripting in the Cloud
Cloud computing platforms provide flexible and elastic computing infrastructure on which a broad range of science and engineering applications are now executed. While cloud proponents tout availability and cost-efficiency as reasons for widespread adoption, in practice cloud resources are often used inefficiently as users lack methods to easily and efficiently scale applications while taking into account the underlying economic models on which the cloud is built. In this talk, I will introduce two projects that aim to enable scalable and cost-efficient use of cloud platforms. First, I will describe Parsl—a parallel scripting library for Python that enables seamless and elastic execution of parallel workflows on clouds, clusters, and supercomputers. Second, I will describe methods for predicting cloud market dynamics and application performance on arbitrary cloud instances. Collectively, these approaches simplify access to on-demand computing infrastructure and improve the efficiency with which applications can be executed on the cloud.
Host: Ian Foster
Kyle Chard is a Research Assistant Professor in the Department of Computer Science at the University of Chicago. He also holds a joint appointment at Argonne National Laboratory. He received his Ph.D. in Computer Science from Victoria University of Wellington, New Zealand in 2011. He is a member of the ACM and IEEE, received the IEEE TCHPC Award for Excellence for Early Career Researchers in HPC, was part of the Globus team that won an R&D100 award, and received the New Zealand Top Achiever Doctoral Scholarship. He co-leads the Globus Labs research group, which focuses on a broad range of research problems in data-intensive computing and research data management. He leads NSF-funded projects related to distributed and parallel computing, scientific reproducibility, research automation, and cost-aware use of cloud infrastructure.