A novel fuzzy QoS based Improved Honey Bee Behavior algorithm for efficient load balancing in cloud

Abstract

Nature inspired algorithm based Load balancing of tasks on virtual machines (VMs) has become an area of greater research interest. Honey Bee Behavior Based Load Balancing (HBB-LB) was introduced to balance the load with a maximum throughput. This approach also balances the priorities of the tasks on the VM to minimize the waiting time of the tasks. However, HBB-LB considers only the VM load for balancing the load, which might not be sufficiently effective. This paper proposes an Improved Honey Bee Behavior Based Load Balancing (IHBB-LB), taking into consideration a few more QoS parameters of VM, such as service response time, availability, reliability, cost and throughput to enhance load balancing. Response time is vital in determining the instant activity of a VM while availability determines available resource and state of VM (idle or active) and Reliability determines the level of trust in a VM. Most importantly, Cost for utilizing a VM and Throughput (capability of VM) are also essential in determining the VM efficiency. But, the inclusion of multiple QoS parameters results in multi-objective optimization problem. As a number of QoS parameters are computed, the Fuzzification of the QoS values was performed through the generated fuzzy rules and multi-objective optimization problem was eliminated. The experiments were performed in terms of makespan, response time, degree of imbalance and the number of tasks migrated and results indicate that the IHBB-LB provides a better level of performance.