Load Balancing Strategy Under Performance Constraints

Cloud computing is a new computing model.Cloud users use it similarly to “hydropower”.Cloud users can pick and play with it.They are not limited by location,area,etc.The price is low.How much service is used to pay for the corresponding fee.With the explosive growth of data,more and more users use cloud services.In the face of massive user requests,cloud computing platforms not only face the problem of uneven resource allocation,waste of resources,and loss of requests,but also large-scale clusters such as data centers.The problem of energy consumption is getting worse.IT industry energy consumption ranks among the highest in the world.For the Internet to develop steadily,reducing energy consumption is a very important task.Therefore,researchers pay more and more attention to the energy consumption of cloud data centers.Aiming at the problem of large energy consumption of existing cloud computing platforms,a load balancing method based on performance/power ratio is proposed.According to the user task request execution process,the overall logical framework of the method is divided into a task processing stage and a resource allocation stage.First,a large number of task simulation experiments are performed to measure the resource consumption thresholds on the physical nodes.Secondly,according to the requirements of the tasks,the processing tasks are classified.In different types of task queues,the tasks are divided into large tasks and common tasks according to the corresponding thresholds.Tasks and small tasks,and complete the task scheduling according to the first-come-first-serving scheduling strategy;Finally,the power consumption of the same type of physical node is predicted based on historical power consumption data,and the performance/power ratio(performance/power consumption)is then constrained.Conditions,select the best physical node to perform the user task request.Simulation experiments show that this method performs better in terms of energy consumption and load balance,effectively guarantees cloud computing platform load balancing,and reduces data center energy consumption to some extent.