Recently, we have witnessed workflows from scientific computing and other data-intensive applications emerging on IaaS clouds. Due to the pay-as-you-go pricing scheme of the cloud, monetary cost and performance become two important optimization factors for running workflows in IaaS clouds. Although many optimization strategies and frameworks have been proposed for the cost and performance optimization problems, little attention has been paid to two important and challenging system issues. First, the optimization strategies and algorithms are usually based on heuristics, limited to specific problem specifications and workflow applications. This is in contrast with users’ different optimization goals (minimizing cost, maximizing performance, etc) and constraints (budget, deadline, etc), as well as evolving application development. Second, the existing designs have almost totally ignored the dynamic features of the cloud, including performance dynamics and price dynamics. Those dynamics are inherent in the cloud, and significantly affect the effectiveness of performance and monetary cost optimizations. In this talk, I will present our recent research efforts in monetary cost optimizations for workflows in the IaaS cloud, addressing the above mentioned two challenges. In the end, I will outline our research agenda in this field.

Bio:
Ms. Amelie Chi Zhou is currently a PhD candidate in School of Computer Engineering of Nanyang Technological University (NTU), under the supervision of Dr. Bingsheng He. Prior to NTU, she obtained her Bachelor’s degree in 2009 and Master’s degree in 2011 both from Beihang University (China). She currently works in the Parallel and Distributed Computing Centre (PDCC) of NTU. Her research interests include cloud computing, distributed systems and big data.