a new method for load balancing in cloud data centers using clustering, requests, and virtual machines (vms)

there are some nodes in cloud systems to meet users’ needs. the volume of the input and processing loads in some of these nodes may sometimes be different from those in other nodes of a cloud, and there may be some imbalance between them as well. many methods have been developed and presented to balance the loads. but the main weaknesses of these methods have been the load imbalance at the moment of the entry of tasks, and lack of attention to the various computational volumes of tasks entering the cloud platform, which will reduce the performance in the cloud, and which will increase the response time, thus ultimately reducing the system performance. using genetic algorithm techniques for producing a generalized genetic algorithm (ga), and using the k-means clustering algorithm for homogeneous clustering by means of virtual machines (vms), we reduced response time, and increased the utilization of processing resources in the cloud platform. first, we created vms with various computational volumes in the cloud platform, and then clustered the tasks entering the cloud and the vms homogeneously with task clusters. afterwards, the generalized genetic algorithm (ga) performed the most precise assignment of task clusters to vm clusters, and then performed the most precise intra cluster assignment at the next level. the method proposed in this paper, on average, has an improvement of at least 152% higher than other algorithms. moreover, the input and experimental data in this paper are derived from the large hadron data center, which is created to reveal the emergence of the universe and the big bang, and which has genuine records.