A Dynamic and Energy Efficient Greedy Scheduling Algorithm for Cloud Data Centers

Rapid progress of mobile and networking technologies resulted in the implementation of an extensive data-centric tasks, which require critical QoS (Quality of Service) and S.L.A (Service Level Agreement) by cloud data centers. This results in critical need for energy efficient task scheduling schemes for data centers. The current algorithms differ from real time cloud scenario and assume that a predefined task schedule is available. To achieve dynamic scheduling of tasks and to improve energy efficiency, we present a model for scheduling the tasks for a cloud data center to scrutinize energy-efficient task scheduling. We formulate the scheduling of tasks to virtual machines (V.M) as an integer-programming problem with the objective of minimizing the energy consumption of the V.M's of the data center and maximizes its residue energy capacities. We prove that the use of a greedy task scheduler confines the S.L.A constraint, minimizes the number of active servers. We conduct extensive experiments on CloudSim tool, with two typical task scheduling algorithms. The experimental results show that our proposed scheme performs better than those algorithms, and can effectively advance the energy utilization of a cloud data center.