Deadline Aware Task Submission and Dynamic Virtual Machine Creation Technique in Cloud Computing

Abstract

The growth of cloud computing has resulted in uneconomic energy consumption, which has negatively impacted the environment through the generation of carbon emissions. One of the most important issues for green cloud environments concerns where to place new virtual machine (VM) requests across physical servers in a way that ensures reduced energy consumption. This paper proposes a deadline based dynamic virtual machine (DBDV) migration algorithm to reduce energy consumption in Cloud datacenters. Virtual machines are classified in three categories: compute intensive task, data intensive task and mix load task. Tasks have been allocated based on similar category of virtual machine. Compute intensive task is assigned to compute intensive virtual machine, data intensive task is assigned to data intensive virtual machine and mix load task is assigned to mix load virtual machine. All tasks are assigned to similar type of virtual machine without classifying virtual machine. The proposed algorithm for virtual machine mapping is based on deadline of task and migration of virtual machine. The proposed algorithm is compared with threshold-based method, PSO and Ant Colony Optimization algorithm and analyzed the result for the energy utilization and number of virtual machine migration as well as completion time. The experimental results show that a DBDV migration algorithm outperforms the existing algorithms in terms of these parameters.