Search

The increasing use of workflows in various fields of science has led to the development of the Workflow as a Service (WaaS) platform, which provides workflow execution as a service to the users. The WaaS platform must use appropriate resource provisioning, sharing, and scheduling, to properly allocate the workflow tasks to the Cloud resources and fulfill the scheduling objectives. In this paper, we have proposed a novel workflow scheduling approach for the WaaS platform that reduces the average makespan of the workflows, improves energy efficiency, and reduces the resource renting cost of the Cloud resources. Our scheduling model consists of containers, virtual machines (VMs), and hosts. We have also proposed a suitable scaling policy for the Cloud resources to improve resource utilization. Extensive simulations with real-world workflows and comparison with some state-of-the-art algorithms show the efficacy of our approach in improving performance and energy efficiency while reducing the monetary cost. [Display omitted] • Multi-criteria scheduling approach for the Workflow as a Service platform. • Optimal MIPS for tasks decides sharing of virtual machine resources with containers. • Resource scaling policy to improve the schedulability of the workflows. • Results show improvement in performance and energy-efficiency while reducing cost. [ABSTRACT FROM AUTHOR]