Towards Enhancing Load-Scheduling and Balancing in Heterogeneous Cloud Environments

Abstract

Load scheduling in heterogeneous cloud computing environments has been a challenge with regard to the waiting times and throughputs. This paper presents a framework that utilizes five agents to bring improvement to the process. The proposed model was simulated using Java Agent Development Environment (JADE) and evaluated using average waiting times and throughput. The dataset was sourced from the Food Concept cloud data center. The simulation varied server scalability from 10 to 50 servers, each with 158 jobs. Lower waiting times of 4.9147s to 5.0692s were achieved demonstrating an improvement over existing techniques in load scheduling. Furthermore, a slight increase in the throughput was realized, showing higher performance of the proposed model. From the results, this study demonstrates that the designed multi-agent load scheduling framework could effectively operate in heterogeneous cloud computing environments, which suggests its potential to decrease task completion times and improve throughput for cloud providers.