Finite Capacity Multi-Server Job Systems: A Simulation Study

Cloud computing has revolutionized how computational resources are accessed and utilized. However, the dynamic nature of the cloud computing environment, which is characterized by a variety of resource types and capabilities presents challenges for managing the workload and ensuring the quality of service. The selection and implementation of queueing policies can have a major impact on the efficiency of the cloud environment, and thus on the quality of service experienced by the end users. Understanding the performance metrics of different queueing policies in cloud computing environments with scalable resource management is essential for both cloud service providers and consumers. In response, our work aims to evaluate the effectiveness of some queueing policies in cloud environments characterized by dynamic resource allocation with a particular emphasis on their dropping probabilities. We proposed a simulation approach that combines the development of an accurate simulation model of a cloud computing environment with adaptable resource management, along with a comprehensive performance analysis of different queueing policies including First-Come-First-Serve and Priority queueing. The result revealed that assigning priority to jobs with longer service times and larger resource demands has a positive impact on small jobs as well.