Managing and Scheduling Approximate Applications to Utilize Renewable Energy in Cloud Computing Datacenters

With the rapid development of cloud computing, the power consumption of datacenters is getting much higher in recent years. Renewable energy becomes promising as the power supply for datacenters since it helps to greatly save the consumption of traditional power resources. In order to improve the utilization of renewable energy in data centers, we proposed several adaptive scheduling algorithms to manage the approximate applications in the datacenter, based on trading off their performance and accuracy. The algorithms are based on different priorities, precisions and running times of the target applications respectively. As the energy generation amount varies with time, the proposed algorithms can follow the trend of energy fluctuation to utilize the renewable energy more efficiently. Evaluation experiment results show that by employing three different kinds of scheduling algorithms, the energy utilization rate can reach 91.9318%, 91.9266% and 91.9225% respectively. It was proved that the adaptive scheduling algorithm could not only help to sufficiently utilize the renewable energy, but also guaranteed a reasonable quality of service for user.