Processing Time Performance Analysis of Scheduling Algorithms for Virtual Machines Placement in Cloud Computing Environment

Abstract

In recent years, a large range of applications have been migrated from traditional computing environments to cloud systems. On the other hand, organizations with existing infrastructure investments leverage the server consolidation in the cloud to serve the upcoming load from various clients around the world. Server consolidation techniques in cloud computing are applied at infrastructure levels and spilled into three main steps; load detection, selection and placement of migrated virtual machines. The placement process can be divided to initial placement and task scheduling while allowing fair load and better resource utilization. The challenge is executing several tasks with the available shared computing resources. Further, the cloud provider should apply a good task strategy depending on the customer’s need. To maintain the best performance of the cloud system, the optimal resource utilization and the shortest completion time for task execution are the essential keys. The heuristics and metaheuristics methods are applied to achieve this goal. In this work, some of the proposed scheduling algorithms will be compared through CloudSim as a simulation tool and through processing time as a performance metric. More specifically, the comparison is made by proposing various simulation scenarios, using the processing time to classify algorithms according to time-shared and space-shared. The evaluation results show that SJF (Short Job First) and FCFS (First Come First Serve) outperforms other baseline algorithms according to different simulations set up.