ISSN 2348-1196 (print) International Journal of Computer Science and Information Technology Research ISSN 2348-120X (online) Vol. 10, Issue 3, pp: (50-61), Month: July - September 2022, Available at: www.researchpublish.com
Load Balancing with Priority Algorithm for Job Replications in Secure Computing Environments Dlaim Alqahtani Saudi Aramco Oil Company Dhahran, Saudi Arabia DOI: https://doi.org/10.5281/zenodo.7056968
Published Date: 07-September-2022
Abstract: Distributed systems in a grid are heterogeneous resources that can be assigned to do major computing tasks using job-scheduling algorithms. Researchers have proposed job-scheduling algorithms in order to improve grid computing from different aspects such as performance, reliability, scalability, fault-tolerance, security and others. In this project, we focus on the limitations of fuzzy-logic based self-adaptive job replication scheduling (FSARS). It is appropriate for fault-tolerant and secure grid job scheduling. Moreover, it addressed the issue of fixed-number job replication scheduling and enhanced it under dynamic security conditions. FSARS can be improved in terms of performance and grid utilization by minimizing average waiting time of jobs and their average response time. Therefore, we propose a new algorithm called load balancing with priority (LBP). This algorithm is implemented for job scheduling in secure grid environment such that it gets the optimal performance compared to FSARS. LBP algorithm has two main characteristics. First, it uses load balancing between hosts that compute jobs and satisfy security conditions. Consequently, this increases grid utilization of its resources and decreases waiting time for jobs. Second, it sorts hosts based on their processing speed and their expected time to compute jobs. As a result, jobs are assigned to hosts with minimum computation time such that host does not exceed average utilization. This algorithm is implemented using MATLAB such that experimental results show improvements in terms of computing performance for job replication scheduling in secure grid environment. Keywords: grid secuirty; job scheduling; grid computing; grid performance; load balancing; grid utilization.
I. INTRODUCTION Grids are geographically distributed systems that act together to form a virtual computing environment in order to perform large computing tasks. These tasks or jobs should be computed such that grid computing is improved from different aspects such as performance, reliability, scalability, fault-tolerance, security and others. Therefore, job scheduling is a significant part in grid computing. In addition, job scheduling has been the concentration of massive amount of research in grid computing [1] [5]. Thus, Job scheduling algorithms have a direct affect in grid computing as they are used for large-scale distributed systems. Researchers developed grid-computing algorithms for job scheduling to achieve optimal results based on different factors to be measured in the grid. They proposed in their research several job scheduling algorithms that can lead to an efficiency of grid resources utilization. Also, they developed job-scheduling algorithms in order to get minimum computation time for jobs and obtain high level of performance in grid environment. In addition, they contributed in security enhancement for job scheduling algorithms [10]. Moreover, researchers developed algorithms such that jobs can tolerate failures by having jobs replications when they are being allocated to grid resources.
Page | 50 Research Publish Journals