Optimal Green Virtual Machine Migration Model

Optimal Green Virtual Machine Migration Model

Noah Sabry (Department of Computing, Faculty of Engineering & Physical Sciences, University of Surrey, Guildford, UK) and Paul Krause (Department of Computing, Faculty of Engineering & Physical Sciences, University of Surrey, Guildford, UK)
DOI: 10.4018/jbdcn.2013070103
OnDemand PDF Download:
$37.50

Abstract

Cloud computing provides the opportunity to migrate virtual machines to “follow-the-green” data centres. That is, to migrate virtual machines between green data centres on the basis of clean energy availability, to mitigate the environmental impact of carbon footprint emissions and energy consumption. The virtual machine migration problem can be modelled to maximize the utility of computing resources or minimizing the cost of using computing resources. However, this would ignore the network energy consumption and its impact on the overall CO2 emissions. Unless this is taken into account the extra data traffic due to migration of data could then cause an increase in brown energy consumption and eventually lead to an unintended increase in carbon footprint emissions. Energy consumption is a key aspect in deploying distributed service in cloud networks within decentralized service delivery architectures. In this paper, the authors address an optimization view of the problem of locating a set of cloud services on a set of sites green data centres managed by a service provider or hybrid cloud computing brokerage. The authors’ goal is to minimize the overall network energy consumption and carbon footprint emissions for accessing the cloud services for any pair of data centres i and j. The authors propose an optimization migration model based on the development of integer linear programming (ILP) models, to identify the leverage of green energy sources with data centres and the energy consumption of migrating VMs.
Article Preview

Virtual machine migration and its application to data centre server farms has been a major focus of research recently (Wood, Shenoy, Venkataramani, & Yousif, 2007; Shrivastava et al., 2011; Meng, Pappas, & Zhang, 2010; Li, Tordsson, & Elmroth, 2011; Tordsson, Montero, Moreno-Vozmediano, & Llorente, 2011; Wubin, Tordsson, & Elmroth, 2012). Sandpiper was proposed in (Wood et al., 2007). Sandpiper uses black and gray-box techniques like monitoring memory, CPU and network utilization to detect hotspots and mitigate them by migrating the VMs to a suitable physical server.

An efficient mechanism for application-aware VM migration was proposed in Shrivastava et al. (2011), through minimizing the network traffic inside data centres.

The work in Meng et al. (2010) formulates a Linear Programme for computing virtual-to-physical machine mappings in the presence of application dependencies. However, their work does not incorporate server-side constraints.

The work in Li et al. (2011) presents a linear integer programming model for dynamic cloud scheduling via migration of VMs across multiple clouds, which offers flexible and elastic price offers.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 13: 2 Issues (2017)
Volume 12: 2 Issues (2016)
Volume 11: 2 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing