DocumentCode :
3439709
Title :
Virtual machine placement for minimizing connection cost in data center networks
Author :
Fukunaga, Takuro ; Hirahara, Shuichi ; Yoshikawa, Hiyori
Author_Institution :
Nat. Inst. of Inf., Tokyo, Japan
fYear :
2015
fDate :
April 26 2015-May 1 2015
Firstpage :
486
Lastpage :
491
Abstract :
Virtualization is a key technology for the efficient operation of massive data centers. To minimize the communication costs among virtual machines (VMs) in a data center network, we formulate an optimization problem for finding efficient VM placements. In this problem, a set of requests is received from customers, where each request is defined as the required number of VMs. The problem seeks to determine those physical machines in the network that host the requested VMs under a capacity constraint such that the number of VMs placed on each physical machine does not exceed that of the available slots. To minimize the load of the networks, for each request, we consider the connection cost of the VM placements, which is defined as the minimum length of networks connecting all physical host machines and the root node. The objective in the problem is to minimize the total connection costs. We present an approximation algorithm for this optimization problem.
Keywords :
approximation theory; computer centres; optimisation; virtual machines; virtualisation; VM placements; approximation algorithm; capacity constraint; communication costs minimization; customer request; data center network; massive data centers; optimization problem; physical host machines; total connection costs minimization; virtual machine placement; virtualization; Approximation algorithms; Approximation methods; Computational modeling; Data communication; Heuristic algorithms; Optimization; Resource management;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Communications Workshops (INFOCOM WKSHPS), 2015 IEEE Conference on
Conference_Location :
Hong Kong
Type :
conf
DOI :
10.1109/INFCOMW.2015.7179432
Filename :
7179432
Link To Document :
بازگشت