Deadline constrained resource provisioning and scheduling algorithm for multiple scientific workflows on waas cloud

Abstract

Workflow is deployed to generate and run computing tasks that are data-dependent, has contributed towards the expedition of scientific development within the scientific community. The workflow make it possible to run scientific jobs that contains a mighty amount of data. These scientific applications are extremely large-size jobs and need extensive computational domain. As a result, they must be deployed in distributed platforms, like cloud domains, in order to obtain an acceptable unit of execution time. newlineWith the growing requirements for scientific workflows execution and the emerging impacts of cloud domains, there is a promising market to offer computational provisioning for processing scientific applications in the computational clouds. As a result, the term Workflow-as-a-Service (WaaS) evolves in tandem with the concept of Everything-as-a-Service (XaaS). This WaaS model enhances the usual Workflow Management System (WMS) execution to accommodate a considerable collection of customers in a low-cost service scenario. Within this scenario, the service that is termed the WaaS framework, should have the power to manage various workflows scheduling and resource provisioning in cloud resources when compared with its single workflow execution of conventional WMS. newlineCloud has transformed a powerful infrastructure for executing scientific workflow applications. Cloud offerings, exclusively Infrastructure as a Service (IaaS) packages, deliver an unlimited pool of computational systems that can be taken as demanded with a pay-as-you-go plan. This pay-per-use scheme completely ignores the requirement for having an infrastructure set-up investment since it can make use of virtualized compute resources that are shared by many. newline