Singapore Institute of Technology
Browse
sensors-23-06155.pdf (4.44 MB)

Prioritized Task-Scheduling Algorithm in Cloud Computing Using Cat Swarm Optimization

Download (4.44 MB)
journal contribution
posted on 2023-07-07, 07:24 authored by Sudheer Mangalampalli, Sangram Keshari Swain, Tulika Chakrabarti, Prasun Chakrabarti, Ganesh Reddy Karri, Martin Margala, Bhuvan Unhelkar, Sivaneasan Bala KrishnanSivaneasan Bala Krishnan

Effective scheduling algorithms are needed in the cloud paradigm to leverage services to customers seamlessly while minimizing the makespan, energy consumption and SLA violations. The ineffective scheduling of resources while not considering the suitability of tasks will affect the quality of service of the cloud provider, and much more energy will be consumed in the running of tasks by the inefficient provisioning of resources, thereby taking an enormous amount of time to process tasks, which affects the makespan. Minimizing SLA violations is an important aspect that needs to be addressed as it impacts the makespans, energy consumption, and also the quality of service in a cloud environment. Many existing studies have solved task-scheduling problems, and those algorithms gave near-optimal solutions from their perspective. In this manuscript, we developed a novel task-scheduling  algorithm that considers the task priorities coming onto the cloud platform, calculates their task VM priorities, and feeds them to the scheduler. Then, the scheduler will choose appropriate tasks for the VMs based on the calculated priorities. To model this scheduling algorithm, we used the cat swarm optimization algorithm, which was inspired by the behavior of cats. It was implemented on the Cloudsim tool and OpenStack cloud platform. Extensive experimentation was carried out using real-time workloads. When compared to the baseline PSO, ACO and RATS-HM approaches and from the results, it is evident that our proposed approach outperforms all of the baseline algorithms in view of the above-mentioned parameters. 

History

Journal/Conference/Book title

Sensors - MDPI

Publication date

2023-07-05

Version

  • Published

Corresponding author

Sudheer Mangalampalli

Usage metrics

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC