Cost effective deadline aware scheduling strategy for workflow applications on virtual machines in cloud computing

Abstract

In Cloud paradigm, resources are provisioned on demand and used for the leased time interval as pay per use basis. Generally, Cloud users use these resources to execute their applications. These applications may be the batch of independent or workflow tasks. The existing researches focus mainly on optimizing the various quality of service (QoS) parameters. In this paper, a Cost Effective Deadline Aware (CEDA) scheduling strategy has been proposed for the workflow tasks/applications in order to optimize the total execution time and economic cost while meeting the deadline and precedence constraints in the workflow. CEDA selects the task which has the highest upward rank and dispatches the selected task to the cheapest instance of virtual machines (VMs) considering the VMs acquisition delay. The performance evaluation of proposed strategy has been carried out by comparing it with ICPCP and ICPCPD2 on standard scientific workflows like Montage, CyberShake, LIGO and SIPHT. The experimental results show that CEDA outperforms among all considered algorithms on almost all parameters under study.

Keywords

• Cloud computing
• Resource provisioning
• Workflows
• Scheduling
• Acquisition delay
• Quality of Service (QoS)