A Scheduling Strategy Of Cloud Computing Data Center Resource With Low Consumption Based On GA

Cloud computing is a kind of on-demand commercial mode with computingresources and storage resources. It is widely used in IT, education, government,finance and other fields. With the expansion of the scale of CCDC (cloud computingdata center) and the dynamic change of demand, energy consumption and resourceutilization become the main restricting factors. We can use effective resourcescheduling strategy to achieve low energy consumption and high resource utilization.The existing CCDC mainly considers the resource utilization, but lacks considerationof energy consumption. How to make the resource scheduling both meet highutilization rate of resources and low energy consumption is a research trend toimprove the performance and economic benefit.This article has expanded the functions during the process of resource schedulingof the existing CCDC from the perspective of low energy consumption. Thenrenaming the function modules as measuring and monitoring platform of energyconsumption, virtual machine manager, energy consumption scheduler according tothe corresponding optimization. And the platform is mainly studied about themeasurement method and the calculation model of energy consumption. From it wecan get the the results based on precision and simplification. Virtual machine managercombines the energy saving technology of close/open. Comparing the energyconsumption from physical machines’ close/open with the energy consumptionwithout loads, we design a judgment of physical machines’ close time. Energyconsumption scheduler communicates with the platform and virtual machine managerin real-time. And the best scheme for resource scheduling is got by genetic algorithm.The improved genetic algorithm is the key to implement the resource schedulingof CCDC. Through this algorithm we can obtained the best mapping scheme betweenvirtual machines and physical machines which meets the given requirements. In thispaper, genetic algorithm is mainly improved by the targets of low energy consumptionand high resource utilization. First of all, studying its tree encoding and allocating theresource based on the setted load threshold in the population initialization with thetarget of high resource utilization. Putting the technique of load balancing to thefitness function and designing the measure formula according to its measurement. Then, with the requirements of the stability and execution efficiency, optimizing thisalgorithm’s operators. Finally, choosing the scheme with the lowest energyconsumption in the conditions of genetic termination.The experiment through CloudSim of cloud computing platform proves that theresource scheduling strategy proposed has a certain extent improvement in terms ofenergy consumption and the effect load balancing. And the improved geneticalgorithm has better stability and execution efficiency.