Content-Based Federated Job Scheduling Algorithm in Cloud Computing
Dinesh Komarasamy, M. Vijayalakshmi · SSRN Electronic Journal · 2016
The resource utilization of the data center has become a big hurdle of commercial cloud service providers due to the rapid expansion and variation of incoming jobs in cloud computing. In the cloud, the end user delivers both deadline and non-deadline based jobs. But, most of the existing scheduling algorithms have autonomously scheduled the deadline and non-deadline based jobs that affected the resource utilization and also violated the SLA policy in terms of deadline. So, the ultimate objective of the proposed work is to execute the deadline and non-deadline based jobs concurrently in a VM. In order to minimize the SLA violations and to improve the resource utilization, this paper proposes a new scheduling technique called Content-based Federated Job Scheduling (CFJS) algorithm in the cloud computing that will deploy in the two-tier VM architecture. In CFJS algorithm, the deadline based jobs are preprocessed based on different job constraints. After preprocessing, the deadline and non-deadline based jobs collect in the correlated job scheduler that are prioritized using the Shortest Deviation First (SDF) method. The prioritized jobs bind with the foreground VM and background VM in the VM that exists in the service provider. These contributions will mitigate the waiting time of the job, increase the resource utilization and avoid starvation. The results are simulated using a Cloudsim toolkit that shows the proposed CFJS algorithm has outperformed the other existing algorithms.