Research Article
A New Resources Provisioning Method Based on QoS
Differentiation and VM Resizing in IaaS
Rongdong Hu,
1
Guangming Liu,
1,2
Jingfei Jiang,
1
and Lixin Wang
1
1
School of Computer, National University of Defense Technology, Changsha 410072, China
2
National Supercomputer Center, Tianjin 300457, China
Correspondence should be addressed to Rongdong Hu; rongdonghu@nudt.edu.cn
Received June ; Accepted July
Academic Editor: Xiaoyu Song
Copyright © Rongdong Hu et al. is is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In order to improve the host energy eciency in IaaS, we proposed an adaptive host resource provisioning method, CoST, which is
based on QoS dierentiation and VM resizing. e control model can adaptively adjust control parameters according to real time
application performance, in order to cope with changes in load. CoST takes advantage of the fact that dierent types of applications
have dierent sensitivity degrees to performance and cost. It places two dierent types of VMs on the same host and dynamically
adjusts their sizes based on the load forecasting and QoS feedback. It not only guarantees the performance dened in SLA, but also
keeps the host running in energy-ecient state. Real Google cluster trace and host power data are used to evaluate the proposed
method. Experimental results show that CoST can provide performance-sensitive application with a steady QoS and simultaneously
speed up the overall processing of performance-tolerant application by ∼%. e host energy eciency is signicantly improved
by ∼%.
1. Introduction
Cloud computing oers near-innite amount of resources
capacity at a competitive rate and allows users to obtain
resources on demand with pay-as-you-go pricing model.
Industry analyst rm IDC predicted that the global cloud
market, including private, public, and hybrid clouds, will hit
billion in and crest at billion by []. e
proliferation of cloud computing has resulted in enormous
amounts of electrical energy consumption.
IaaS (infrastructure as a service), as one important form
of cloud computing, mainly leverages the virtualization tech-
nology to create multiple VMs (virtual machines) on a phys-
ical host and can support rapid deployment of large-scale
applications []. Cloud providers can reduce power con-
sumption by consolidating various applications into a fewer
number of physical hosts and switching idle hosts to low-
power modes. However, virtualization also creates a new
problem. e applications performance relies on eective
management of VM capacity. One essential requirement of
cloud computing is providing steady QoS dened in terms
of SLA (service level agreements). SLA violation will bring
economic penalties to cloud providers. erefore, they always
strive to ensure the agreed performance of individual VM.
Consequently, performance control and power manage-
ment are two major research issues in modern data centers.
But they are in conict. Seeking a good tradeo between
power consumption and applications performance is crucial
for cloud providers.
e focus of this work is on the energy eciency of IaaS
cloud data center. Two typical applications are considered:
performance-sensitive application (sVM) and performance-
tolerant application (tVM). CoST places them on the same
host and dynamically adjust their sizes based on the load
forecasting. From the point of sVM, tVM is an elastic resource
pool. e requirement of sVM dominates the host resource
allocation. CoST not only guarantees the QoS, but also keeps
the host running in energy-ecient state. Experiments with
real trace data in CloudSim show that CoST can speed up
the overall progress of tVM and improve the host energy
eciency signicantly.
e rest of this paper is organized as follows. Section
describes the background and motivation. Section discusses
our analysis results, the detailed design and implementation
Hindawi Publishing Corporation
Mathematical Problems in Engineering
Volume 2015, Article ID 215147, 9 pages
http://dx.doi.org/10.1155/2015/215147