A Novel Artificial Bee Colony Approach of Live Virtual Machine Migration Policy Using Bayes Theorem

Gaochao Xu,Liang Hu,Yan Ding,Jia Zhao,Xiaodong Fu

doi:10.1155/2013/369209

Abstract

Green cloud data center has become a research hotspot of virtualized cloud computing architecture. Since live virtual machine (VM) migration technology is widely used and studied in cloud computing, we have focused on the VM placement selection of live migration for power saving. We present a novel heuristic approach which is called PS-ABC. Its algorithm includes two parts. One is that it combines the artificial bee colony (ABC) idea with the uniform random initialization idea, the binary search idea, and Boltzmann selection policy to achieve an improved ABC-based approach with better global exploration's ability and local exploitation's ability. The other one is that it uses the Bayes theorem to further optimize the improved ABC-based process to faster get the final optimal solution. As a result, the whole approach achieves a longer-term efficient optimization for power saving. The experimental results demonstrate that PS-ABC evidently reduces the total incremental power consumption and better protects the performance of VM running and migrating compared with the existing research. It makes the result of live VM migration more high-effective and meaningful.

Highlights

virtual machine (VM) technology [1, 2], one of the most important technologies in cloud computing, is a way to implementing cloud computing such as Infrastructure as a Service (IaaS) [3] architecture and for the embody of the cloud computing idea, whereas live VM migration technology, which is widely used for the maintenance management in virtualized cloud computing data centers, is the representative of the VM technologies
During the course of the study, we have found that the search process of PS-artificial bee colony (ABC) can be regarded as the repetitive randomized trials since it has the characters of randomized trials such as the nondeterminacy of results, the repeatability of trials, and the diversity of results and so forth
In the improved ABC-based approach, we introduce the binary search idea into the classic ABC approach to achieve the uniform random initialization and to make the whole PS-ABC approach have a better global search potential and capacity at the very beginning

Summary

Introduction

VM technology [1, 2], one of the most important technologies in cloud computing, is a way to implementing cloud computing such as Infrastructure as a Service (IaaS) [3] architecture and for the embody of the cloud computing idea, whereas live VM migration technology, which is widely used for the maintenance management in virtualized cloud computing data centers, is the representative of the VM technologies. On the background of low-carbon world and cloud computing era, researchers have proposed the field of green cloud computing based on cloud computing and virtualization. It aims at reducing power consumption in cloud computing data centers. Only one target host is most suitable for the VM in order to minimize the total incremental power consumption in cloud data center. To achieve green cloud data center, in the search direction of VM provisioning the problem is similar to that of live VM migration. Both are to find out the optimal target host

Objectives

Methods

Results

Conclusion