An Energy-Efficient Resource Scheduling Algorithm for Cloud Computing based on Resource Equivalence Optimization

An Energy-Efficient Resource Scheduling Algorithm for Cloud Computing based on Resource Equivalence Optimization

Li Mao (Department of Computer Science, Guang Dong Police Officer College, Guangzhou, China & School of Computer Science and Engineering, South China University of Technology, Guangzhou, China), De Yu Qi (School of Computer Science and Engineering, South China University of Technology, Guangzhou, China), Wei Wei Lin (School of Computer Science and Engineering, South China University of Technology, Guangzhou, China), Bo Liu (School of Computer, South China Normal University, Guangzhou, China) and Ye Da Li (Guangdong Engineering Polytechnic, Guangzhou, China)
Copyright: © 2016 |Pages: 15
DOI: 10.4018/IJGHPC.2016040103
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

With the rapid growth of energy consumption in global data centers and IT systems, energy optimization has become an important issue to be solved in cloud data center. By introducing heterogeneous energy constraints of heterogeneous physical servers in cloud computing, an energy-efficient resource scheduling model for heterogeneous physical servers based on constraint satisfaction problems is presented. The method of model solving based on resource equivalence optimization is proposed, in which the resources in the same class are pruning treatment when allocating resource so as to reduce the solution space of the resource allocation model and speed up the model solution. Experimental results show that, compared with DynamicPower and MinPM, the proposed algorithm (EqPower) not only improves the performance of resource allocation, but also reduces energy consumption of cloud data center.
Article Preview

1. Introduction

In recent years, with the rapid increasing of power consumption in global data centers and IT systems, high energy-consumption gradually becomes a prominent issue. Especially, with the arrival of the cloud computing era (Chen & Zheng, 2009; Foster, Zhao, Raicu, & Lu, 2008; W.-W. Lin & Qi, 2012), more and more computing and storage resources pile up in the cloud centers, which brings great challenges to the management of the energy consumption (Foster, Zhao, et al., 2008; Zhao-Hui & Qin-Ming, 2012). According to statistics(Hooper, 2008), about 6000 data centers consumed approximately 61 billion kilowatt-hours of electricity costing $4.5 billion in the United States in 2006, which is more than the total energy consumption of all televisions in the United States. The data from the United States energy department shows that the energy consumption in data centers accounts for 1.5% of all the energy consumption in USA, and the demand for electricity in date centers is still growing at an annual rate of 12%. By 2011, the data centers consume 100 billion kilowatt-hours of electricity costing about $7.4 billion per year. In addition, the market research to all the electricity cost in the whole world made by IDC, the International Data Corporation shows that the global enterprises would cost about $40 billion a year on the energy consumption. High energy consumption in data center not only wastes too much energy and makes the running systems not stable, but also brings negative effects to the environment of nature. The federal agency of USA has pointed out that high energy consumption makes serious influences to the air quality, national security, climate changes and the reliability of power grid, and etc. Therefore, the energy consumption optimization has become an important issue to be solved urgently in current cloud data centers.

Based on the important virtualization technologies, cloud computing has become the delivery of the resources and supported the pay-as-you-go model. In a virtualized cloud data center, thousands of phisical machines (PMs) are composited of a pool of computing resources to host the various virtual machines (VMs) that provide desirable environments for any users. VM migration is an efficient tool for resource provisioning, by dynamically rearranging the previous placement (Kangkang Li, Zheng, & Wu, 2013). Taking the advantage of VM migration, VMs save costs by reducing the need for PMs, and lower the power and cooling demands for more efficiently using hardware.

The energy-efficient scheduling VMs aims at sharing resources (CPU etc.) fairly among multiple VMs, assigning VMs to PMs in accordance with the algorithm, which minimizes the number of PMs used. In practice, the service providers of cloud data centers own all the details of the PMs in the cloud and different hardware requirements of VMs from the users, they can place VMs to proper physical machines to minimize the energy and thus maximize their profit(Ding, Qin, Liu, & Wang, 2015). It is important that an energy-efficient VM scheduling algorithm should consider the heterogeneous characteristics of PMs, such as the number of cores and the size of memory, because the servers in the cloud data center are always heterogeneous.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing