With the growth of cloud computing, large scale data centers have become common in the computing industry, and there has been a significant increase in energy consumption at these data centers, which thus becomes a key issue to address. As most of the time a data center remains underutilized, a significant amount of energy can be conserved by migrating virtual machines (VM) running on underutilized machines to other machines and hibernating such underutilized machines. This paper aims to design such a strategy for energy-efficient cloud data centers. It makes use of historical traffic data from data centers and uses a service request prediction model which enables the identification of the number of active servers required at a given moment, thus making possible the hibernation of underutilized servers. The simulation results indicate that this approach brings about a significant amount of energy conservation.Index Terms-cloud computing, energy conservation, service level agreement, historical data I. INTRODUCTIONWith the growing needs of on-demand access [1] and high computational power, cloud computing has emerged as a new business model [2]. Cloud vendors rely on large-scale computing infrastructures that consume enormous amounts of electrical power at the scale of megawatts. Therefore, attention is required towards minimizing energy requirement. With time, the operational cost of cloud computing infrastructure exceeds the initial cost of the infrastructure. Research shows that running a single 300-watt server during a year can cost about $338, and more importantly, can emit as much as 1,300 kg of CO 2 [1]. It is estimated that in 2006, the cost of electricity consumed by IT infrastructure in the US was around $4.5 billion US, which came to about 1.5% of the total US energy consumption that year; these figures are expected to double by 2011. By 2015, the costs of operations, of which the cost of electrical power is an important part, will cross the initial cost of IT infrastructure or hardware [3].Taking these facts into consideration, it can be deduced that there is a growing need to contain the energy consumed by large and complex IT systems. Not only would this result in greener, more energy-efficient IT systems, but would also help in improving the financial performance of the organizations owning and using these systems. Generally, all the computing resources of a large-scale cluster keep running 365 × 24, regardless of the actual time-varying computing requirements, and this may result in lower resource utilization and additional needless power consumption even when the computing demand is not large enough. In some embedded systems, energy consumption determines the lifetime of the equipment. A server cluster can be termed as QoS-aware if it reduces energy consumption while maintaining its Service Level Agreement (SLA).

No abstract

—With the growth of cloud computing, large scale data centers have become common in the computing industry, and there has been a significant increase in energy consumption at these data centers, which thus becomes a key issue to address. As most of the time a data center remains underutilized, a significant amount of energy can be conserved by migrating virtual machines (VM) running on underutilized machines to other machines and hibernating such underutilized machines. This paper aims to design such a strategy for energy‐efficient cloud data centers. It makes use of historical traffic data from data centers and uses a service request prediction model which enables the identification of the number of active servers required at a given moment, thus making possible the hibernation of underutilized servers. The simulation results indicate that this approach brings about a significant amount of energy conservation.