Simulation Of Resource Scheduling Algorithm Of Cloud Data Centers

In order to meet the current massive information processing needs and increasing computing power requirements,the scale of cloud computing continues to expand.As the infrastructure of cloud computing,the data center has also developed rapidly.The ever-larger data centers have generated huge energy consumption.At the same time,the current resource scheduling algorithms in the data center cannot fully balance energy consumption and other performance requirements,which seriously restricts the development of cloud computing.Resource scheduling optimization has become a hot and difficult point in current research.In this paper,resource scheduling optimization is carried out from two aspects: virtual machine placement and task assignment,and the energy consumption model based on simulation technology is used for evaluation.The main research contents are as follows:(1)A multivariate nonlinear simulation energy consumption model is proposed.The model is based on CPU and memory usage,and considers the impact of CPU utilization on memory,which improves the accuracy of simulation energy consumption assessment.The prediction of the model is compared with the multivariate linear model,the CloudSim linear model and the CloudSim segmentation model.The results show that the proposed multivariate nonlinear energy consumption model has higher accuracy.The model is used to rewrite the energy consumption module in the CloudSim simulation platform,which is applied to the energy consumption evaluation of subsequent simulation experiments.(2)A virtual machine placement algorithm based on BFD and ant colony optimization is proposed.The algorithm applies the improved BFD algorithm to the pheromone initialization process of ant colony algorithm,avoids blind search of ant colony algorithm,and designs energy optimization.The state transition rule and the updating method of global pheromone,in order to avoid the quality of service caused by excessive resource utilization,an elite policy based on the resource utilization threshold is set in the updating method of global pheromone.Then the algorithm is implemented in CloudSim simulation platform,and compared with BFD algorithm,improved BFD algorithm and improved ant colony algorithm.The results show that the proposed algorithm has better optimization effects in both energy consumption and service quality.(3)A task assignment algorithm based on Max-Min and Genetic Algorithm is proposed.The algorithm applies Genetic Algorithm to the task assignment process,and uses chromosome coding to represent the mapping relationship between tasks and virtual machines,uses the combination of improved max-min and stochastic initialization to produce the initial generation population.The fitness function aiming at reducing energy consumption and shortening the execution time is designed,and the load difference threshold is added to the genetic operator to avoid serious load imbalance.Then the algorithm is implemented in the CloudSim simulation platform,and compared with the standard Max-Min and the standard Genetic Algorithm.The results show that the proposed algorithm has obvious advantages in reducing energy consumption and shortening execution time,and has better load equilibrium.