Stratus cluster scheduler: enhancing cloud efficiency and reducing costs in VM deployment

• Main Author: Kho, Marcus Han Kiat
• Other Authors: Tang Xueyan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Cloud computing
• Online Access: https://hdl.handle.net/10356/181132
• Institution: Nanyang Technological University

The report investigates the efficiency of the Stratus cluster scheduler in optimizing resource allocation within a public cloud environment and is focused on Virtual Machine deployment. In a public cloud, when there is an increasing demand for cloud services, traditional assignments-a VM for every task-suffer from huge underutilization of resources, with high costs. Using CloudSim Plus with Google trace data (2011v2), the study evaluates the performance of Stratus against a baseline configuration that shows huge improvements in VM CPU utilization, task execution times, and total cost efficiency. As obtained from the results, it was shown that Stratus consolidated the tasks into a small number of VMs, hence reducing the overall up-time of the VMs by about 99% as compared to the baseline, which improves resource utilization efficiency and reduces operational costs. These findings point to the potential of Stratus to address critical inefficiencies in managing resources on the cloud. Future research directions are outlined, emphasizing the investigation of further scalability of Virtual Cluster scheduling and the integration of emerging technologies in order to continue the aggregation of optimization performed at cloud services.