International Journal of Science, Engineering and TechnologyAuthors: Dr. M.D. Vimalapriya, Ms. Nisha
Abstract: Queuing theory forms the core mathematics behind modeling, analyzing, and optimizing cloud computing systems. In this work, the researchers undertake a rigorous computational analysis of the applications of queuing theory on clouds with emphasis on some important performance parameters such as response time, resource utilization, and throughput. A general open Jackson network of queues that incorporates the multilevel nature of modern cloud data center networks, i.e., entrance servers, server clusters, and database tiers, is proposed for this purpose. This is accomplished through discrete-event simulation using MATLAB and CloudSim packages. The findings show that the proposed queueing model performs considerably well with respect to the response time (49.5% improvement), and resource utilization (30.08%) when compared to other methods. Also, it highlights the impact of some crucial factors such as workloads' nature, servers' scalability, and catastrophe mechanism on the system stability and traffic control.
DOI: http://doi.org/10.5281/zenodo.22724567
