A Novel Resource Optimization Model on Realtime Cloud Computing using Bayesian Estimation

Abstract

As the size of the cloud computing resources and services increases, it is difficult to handle load balancing due to computational cost and time. Since, most of the cloud service providers have their own type, type and price policies for computing resources, including other service features. The load balance between cloud resources ensures an efficient utilization of the physical infrastructure while minimizing runtime. Load balance can improve quality of service (QoS) measurements, including response time, cost, performance and use of resources. In this work, a novel load balancing algorithm is implemented to improve the cloud service load balancing. In order to optimize the delivery of cloud services, the load balance is important between virtual machines at minimum paid costs and overall service delivery time. In order to improve the scheduling process of load-balancing in the cloud environment, many traditional models are used to optimize the load balance. However, the main problem to the cloud service provider's is optimizing cloud service parameters such as reliability, flexibility, time limits and the task refusal rate. A dynamic algorithm is required for the cloud service provider to plan work which will reduce time while increasing the cloud resources use ratio and comply with the user's specific QoS parameters. The proposed bayesian scoring function based PSO model  is based on hybridizing heuristic techniques with metaheuristic algorithm in order to achieve its optimum performance in the load balancing process. Experimental results proved that the present load-balancing model has better performance than the traditional load balancing approaches on various cloud resources.