Evaluating the performance of load balancing algorithm for heterogeneous cloudlets using HDDB algorithm

Load balancing is the major concern in cloud computing where number of requests have to be handled by cloud resources. The load balancing techniques distribute the workloads among the computing resources and manage the cloud resources optimally. The load balancing algorithms seek to balance the system load by moving workloads from the overloaded resources to underloaded resources in order to ensure the balancing of overall workload in the cloud environment. The aim of this study is to allocate virtual machines to the best-suited hosts based on CPU availability and host membership value using Hybrid Dynamic Degree Balance (HDDB) algorithm. The proposed scheduling technique is utilizing two algorithms namely Dynamic Degree Balance CPU Based (D2B_CPU based) and Dynamic Degree Balanced Membership based (D2B_Membership) to present a hybrid technique which is capable of balancing the workload optimally. The suggested algorithm HDDB has been tested using the CloudSim simulation tool. To verify the performance of proposed hybrid algorithm, performance metrics are used in contact with turnaround time of cloudlets, execution cost, throughput time, degree of imbalance, CPU utilization, bandwidth utilization and memory utilization. The results reveal a considerable improvement in performance of the hybrid load balancing method when compared to other existing algorithms.