Adaptive Workload Scheduling Design in Parallel Computational Framework

Abstract

In the recent years, many researchers have showed a great deal of interest to improve the scheduling of workload in the cloud platforms. On the other hand, to carry out the execution of the scientific workloads in the cloud environment, it consumes much time and is expensive; hence, it is neither time efficient nor cost-efficient. Due to this reason, many research studies have been carried, by which the researchers tend to reduce the processing time and make a cost-efficient method as the users are charged based on the usage. Very few studies have been done to optimize the cost with processing time and energy parameters together in order to meet the Service Level Agreement (SLA) and Quality of Service (QoS) of the workload task. Hence, in this paper, we present an Adaptive Workload Scheduling (AWS) model that ensures the Task Level SLA prerequisites in a heterogeneous distributed-computing environment. This AWS- model approach reduces the amount of energy and time needed to execute a given workloads. Cybershake scientific workload has been utilized for the studying proposed AWS model. When model was compared with the standard workload scheduling approach, our model reduced the consumption of cost, energy and time.