| With the rapid development of cloud computing,the pollution and impact on the environment has become increasingly serious what is caused by the constantly expanding of datacenter's scale.It has become a new trend to use the renewable as energy supply of datacenters.This paper's research object is based on datacenters powered by renewable energy,studying themodel and algorithms of resource and energy management.It is the premise of the scheduling problem about how to predict the energy consumption of renewable energy sources and data center energy consumption modeling.Two energy consumption modeling methods are proposed in this paper,in order to put forward the two kinds of datacenters energy consumption scheduling algorithms by giving the prediction results of solar power generation.Finally,two forecasting models are established,which are multiple linear regression and artificial neural network,via finding out the factors that affect the solar radiation intensity.According to research the factors that affect the energy consumption of data center,the CPU energy consumption model and energy consumption model was established,since the energy consumption model is the basis of the cloud computing resource scheduling algorithm.Finally,this paper proposed two kinds of cloud data center scheduling algorithms.Firstly,we proposed to research on renewable-energy-aware approximate application scheduling algorithm,for three different objectives we designed three kinds of adaptive scheduling algorithms respectively,including the priority weighting algorithm,the high precision energy consumption weighting algorithm,and the running time weighting algorithm.Then we proposed to research on renewable-energy-aware GA(Genetic Algorithm)-based virtual machine scheduling algorithm.The simulation results show that the algorithm can make full use of the solar energy generated by the running of the data center using CloudSim cloud computing simulation software. |
PDF Full Text Request