A cluster medoid approach for cloud task scheduling

Article type: Research Article

Authors: Raju, Y. Home Prasanna^{a; *} | Devarakonda, Nagaraju^b

Affiliations: [a] Department of CSE, Acharya Nagarjuna University, Guntur, India | [b] School of Computer Science and Engineering, VIT-AP University, Amaravati, India

Correspondence: [*] Corresponding author: Y. Home Prasanna Raju, Department of CSE, Acharya Nagarjuna University, Guntur, A.P., India. E-mail: yhprasannaraju@gmail.com.

Abstract: One of the familiar distributed technologies for sharing computing resources through internet is a cloud computing technology. One need not setup all computing resources on their own to design their applications. They can own as much they want by requesting computing resources through net. These resources are shared between users upon request by properly scheduling tasks in cloud. The process of scheduling tasks is to be optimized to share the resources very fast. The paper proposes a cluster medoid based task scheduling technique KMPS (K-medoid particle swarm approach) for minimizing the makespan. KMPS uses the merits of both Particle Swarm Optimization (PSO) and k-medoid approaches with added weights concept. Experimental results have shown that KMPS has optimized the results of make span and it is most suitable one for cloud computing.

Keywords: Cloud computing, k-medoids, makespan, PSO, task scheduling

DOI: 10.3233/KES-210053

Journal: International Journal of Knowledge-based and Intelligent Engineering Systems, vol. 25, no. 1, pp. 65-73, 2021