TY - JOUR
T1 - An Intelligent Optimization Method for Optimal Virtual Machine Allocation in Cloud Data Centers
AU - Zhang, Peiyun
AU - Zhou, Meng Chu
AU - Wang, Xuelei
N1 - Funding Information:
Manuscript received October 29, 2018; revised February 5, 2019, July 13, 2019, and December 11, 2019; accepted February 9, 2020. Date of publication March 9, 2020; date of current version October 6, 2020. This article was recommended for publication by Associate Editor G. Hu and Editor F.-T. Cheng upon evaluation of the reviewers’ comments. This work was supported in part by the National Natural Science Foundation of China under Grant 61872006 and Grant 61472005 and in part by the China Education and Research Network (CERNET) Innovation Project under Grant NGII20160207. Its earlier version was presented at the IEEE CASE 2017, titled “Multi-user Multi-provider Resource Allocation in Cloud Computing.” Peiyun Zhang and Xuelei Wang are with the School of Computer and Information, Anhui Normal University, Wuhu 241002, China (e-mail: zpyanu@ahnu.edu.cn; wxl_raymond@163.com).
Publisher Copyright:
© 2004-2012 IEEE.
PY - 2020/10
Y1 - 2020/10
N2 - A cloud computing paradigm has quickly developed and been applied widely for more than ten years. In a cloud data center, cloud service providers offer many kinds of cloud services, such as virtual machines (VMs), to users. How to achieve the optimized allocation of VMs for users to satisfy the requirements of both users and providers is an important problem. To make full use of VMs for providers and ensure low makespan of user tasks, we formulate an optimal allocation model of VMs and develop an improved differential evolution (IDE) method to solve this optimization problem, given a batch of user tasks. We compare the proposed method with several existing methods, such as round-robin (RR), min-min, and differential evolution. The experimental results show that it can more efficiently decrease the cost of cloud service providers while achieving lower makespan of user tasks than its three peers. Note to Practitioners-VM allocation is one of the challenging problems in cloud computing systems, especially when user task makespan and cost of cloud service providers have to be considered together. We propose an IDE approach to solve this problem. To show its performance, this article compares the commonly used methods, i.e., RR and min-min, as well as the classic differential evolution method. A cloud simulation platform called CloudSim is used to test these methods. The experimental results show that the proposed one can well outperform its compared ones, and its VM allocation results can achieve the highest satisfaction of both users and providers. The proposed method can be readily applicable to industrial cloud computing systems.
AB - A cloud computing paradigm has quickly developed and been applied widely for more than ten years. In a cloud data center, cloud service providers offer many kinds of cloud services, such as virtual machines (VMs), to users. How to achieve the optimized allocation of VMs for users to satisfy the requirements of both users and providers is an important problem. To make full use of VMs for providers and ensure low makespan of user tasks, we formulate an optimal allocation model of VMs and develop an improved differential evolution (IDE) method to solve this optimization problem, given a batch of user tasks. We compare the proposed method with several existing methods, such as round-robin (RR), min-min, and differential evolution. The experimental results show that it can more efficiently decrease the cost of cloud service providers while achieving lower makespan of user tasks than its three peers. Note to Practitioners-VM allocation is one of the challenging problems in cloud computing systems, especially when user task makespan and cost of cloud service providers have to be considered together. We propose an IDE approach to solve this problem. To show its performance, this article compares the commonly used methods, i.e., RR and min-min, as well as the classic differential evolution method. A cloud simulation platform called CloudSim is used to test these methods. The experimental results show that the proposed one can well outperform its compared ones, and its VM allocation results can achieve the highest satisfaction of both users and providers. The proposed method can be readily applicable to industrial cloud computing systems.
KW - Cloud computing
KW - improved differential evolution (IDE)
KW - intelligent optimization
KW - virtual machine allocation
UR - http://www.scopus.com/inward/record.url?scp=85085491260&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85085491260&partnerID=8YFLogxK
U2 - 10.1109/TASE.2020.2975225
DO - 10.1109/TASE.2020.2975225
M3 - Article
AN - SCOPUS:85085491260
SN - 1545-5955
VL - 17
SP - 1725
EP - 1735
JO - IEEE Transactions on Automation Science and Engineering
JF - IEEE Transactions on Automation Science and Engineering
IS - 4
M1 - 9027813
ER -