Energy-efficient cloud resource management

We propose a resource management framework that reduces energy consumption in cloud data centers. The proposed framework predicts the number of virtual machine requests along with their amounts of CPU and memory resources, provides accurate estimations of the number of needed physical machines, and reduces energy consumption by putting to sleep unneeded physical machines. Our framework is based on real Google traces collected over a 29-day period from a Google cluster containing over 12,500 physical machines. Using this Google data, we show that our proposed framework makes substantial energy savings.