A Hybrid Multiple Parallel Queuing Model to Enhance QoS in Cloud Computing

Abstract

Among the different QoS metrics and parameters considered in cloud computing are the waiting time of cloud tasks, execution time of tasks in VM's, and the utilization rate of servers. The proposed model was developed to overcome some of the pitfalls in the existing systems among which are sub-optimal markdown in the queue length, waiting time, response time, and server utilization rate. The proposed model contemplates on the enhancement of these metrics using a Hybrid Multiple Parallel Queuing approach with a joint implementation of M/M/1: ∞ and M/M/s: N/FCFS to achieve the desired objectives. A neoteric set of mathematical equations have been formulated to validate the efficiency and performance of the hybrid queuing model. The results have been validated with reference to the workload traces of Bit Brains infrastructure provider. The results obtained indicate the significant reduction in the queue length by 60.93 percent, waiting time in the queue by 73.85 percent, and total response time by 97.51%.