J Grid Computing https://doi.org/10.1007/s10723-019-09490-2 An Energy Efficient Algorithm for Workflow Scheduling in IaaS Cloud Vishakha Singh · Indrajeet Gupta · Prasanta K. Jana Received: 17 July 2018 / Accepted: 16 August 2019 © Springer Nature B.V. 2019 Abstract Energy efficient workflow scheduling is the demand of the present time’s computing platforms such as an infrastructure-as-a-service (IaaS) cloud. An appreciable amount of energy can be saved if a dynamic voltage scaling (DVS) enabled environ- ment is considered. But it is important to decrease makespan of a schedule as well, so that it may not extend beyond the deadline specified by the cloud user. In this paper, we propose a workflow schedul- ing algorithm which is inspired from hybrid chemical reaction optimization (HCRO) algorithm. The pro- posed scheme is shown to be energy efficient. Apart from this, it is also shown to minimize makespan. We refer the proposed approach as energy efficient workflow scheduling (EEWS) algorithm. The EEWS is introduced with a novel measure to determine the amount of energy which can be conserved by consid- ering a DVS-enabled environment. Through simula- tions on a variety of scientific workflow applications, we demonstrate that the proposed scheme performs better than the existing algorithms such as HCRO and V. Singh · P. K. Jana Department of Computer Science & Engineering, Indian Institute of Technology (ISM), Dhanbad 826004, India e-mail: vs.make.a.vish@gmail.com Prasanta K. Jana e-mail: prasantajana@yahoo.com I. Gupta () Department of Computer Science Engineering, Bennett University, Greater Noida, 201310, India e-mail: indrajeet7830@gmail.com multiple priority queues genetic algorithm (MPQGA) in terms of various performance metrics including makespan and the amount of energy conserved. The significance of the proposed algorithm is also judged through the analysis of variance (ANOVA) test and its subsequent LSD analysis. Keywords Workflow scheduling · Energy conservation · Chemical reaction optimization · Makespan · Cloud 1 Introduction Workflow scheduling in cloud computing continues to attract the attention of research fraternity, as it lever- ages the full strength of distributed computing [1–10]. A real world workflow application consists of a set of a large number of interdependent tasks. Such workflows can be represented as directed acyclic graphs (DAGs), which are executed on infrastructure as a service (IaaS) cloud to run the applications. An IaaS cloud is a deployment model used in cloud computing which provides computational resources to the users for exe- cuting their applications. Scheduling of workflows is the major concern for the cloud service provider (CSP) which furnishes IaaS cloud resources to its users on the basis of pay-as-you-go model. Workflow scheduling consists of two phases: 1) determining the execution order of tasks without violating any of the