Scheduling of scientific workflows in cloud with replication

Abstract

Cloud infrastructure is rapidly scalable, accessible and also it is a pay as we use model. Due to the elastic nature of cloud, the resource provisioning can be done dynamically according to the requirements of the workflow application. We aim at executing the scientific workflows within the user-defined deadline, with the smaller possible cost by reducing the total execution time of applications. The execution time of the task is reduced by proper provisioning and scheduling of all tasks in the scientific workflows. The proposed algorithm RIPCP (Replication based IC-PCP) uses the principle of IC-PCP algorithm for finding the Partial Critical Path (PCP) of the tasks and includes the task replication for better make span. Three different Scientific Workflows are simulated in different characteristics of virtual machines to demonstrate that the proposed algorithm reduces the total execution time of the applications.