Balance Resource Utilization (BRU) Approach for the Dynamic Load Balancing in Cloud Environment by Using AR Prediction Model

One of the major challenges for the cloud provider is the efficient utilization of the physical resources. To achieve this, this paper proposed a Balance Resource Utilization (BRU) approach that not only minimizes the resource leakage but also increases the resource utilization and optimize the system performance. The proposed approach consider two resources i.e., CPU and memory, as decision metrics for load balancing and use three thresholds named lower threshold, upper threshold and warning threshold to define underloaded, overloaded and warning situations, respectively. The main concept of this approach is to place VM to the PM, where resource requirement of the VM and resource utilization of the PM are complements to each other. To evade unnecessary migrations due to the temporary peak load AR time series prediction model is used. The authors' approach treats load balancing problem from the practical perspective and implemented in OpenStack cloud with KVM hypervisor. Moreover, proposed approach resolve the issue of VM migration in the heterogeneous environment.