Fueled by promises of enhanced util

Fueled by promises of enhanced utilization, improved
efficiency, and increased flexibility, system administrators have
expanded their IT architectures to include a large number
of virtualized systems [1]. Further, virtulization is one of
the cornerstone technologies that makes utility computing
platforms such as cloud computing a reality. For example
the industry leader Amazon provides computing as a service
through it’s Xen virtualization based EC2 platform. There have
been significant studies on the performance and optimization of
virtual machines (VMs), including their power consumption in
performing different types of tasks. Research has been done
on reducing power consumption of data-centers through the
placing of energy consuming jobs/VMs in strategic cooling
locations, energy reduction through job and VM consolidation,
and energy aware scheduling [2] [3] [4] [5] [6] [7] [8]. Other
pioneering works have explored solutions to meter or cap the
energy consumption of virtual machines running in a cloud
environment [9] [10]. Work has also been done on power
consumed during a VM migration [11]. Further, the energy
consumption implications of VMs used in server consolidation
has been discussed in [12]