Information Technology Reference
In-Depth Information
Energy-Aware Multi-Agent Server Consolidation
in Federated Clouds
Alessandro Ferreira Leite and Alba Cristina Magalhaes Alves de Melo
Department of Computer Science
University of Brasilia, Brasilia, Brazil
Abstract. In this paper, we propose and evaluate a server consolida-
tion approach for ecient power management in virtualized federated
Data Centers. The main goal of our approach is to reduce power con-
sumption, trying to meet QoS requirements with limited energy defined
by a third party agent. In our model, we address application workload
considering the costs due to turning servers on/off and Virtual Machine
migrations in same Data Center and between different Data Centers.
Our simulation results with 2 data centers and 400 simultaneous Virtual
Machines show that our approach is able to reduce more than 50% of
energy consumption, while still meeting the QoS requirements.
1
Introduction
Cloud Computing is a recent paradigm for provision of computing infrastructure,
platform and/or software. This paradigm shifts the location of these components
to the Internet in order to reduce costs associated with resource management
(hardware and software) [10].
Cloud Computing is gaining popularity since it helps companies to reduce
costs and carbon footprint. Usually, services are executed in big Data Centers
containing a large number of computing nodes. The energy requirements of the
whole Data Center have a high impact on the total operation costs [11], which can
be over 60% of the peak load [8, 15]. Therefore, reducing the energy consumption
without sacrificing Quality-of-Service (QoS) requirements is an important issue.
In Cloud Computing, Data Centers usually employ virtualization techniques
to provide computing resources as utilities and Virtual Machine (VM) technolo-
gies for server consolidation. Server consolidation is the process of gathering
several virtual machines into a single physical server. It aims at minimizing the
number of physical servers required to host a group of Virtual Machines.
Many studies have been conducted to provide power reduction and some of
them are based in server consolidation [1]. However, server consolidation in Cloud
Computing can introduce some diculties such as: (i) the Cloud environment
must usually provide Quality of Service (QoS) guarantees, normally defined in
terms of Service Level Agreements (SLA); (ii) it is common to occur dynamic
changes of the incoming requests rate; (iii) the usage pattern of the resources
is often unpredictable and (iv) different users have distinct preferences.
