Delay-based workflow scheduling for cost optimization in heterogeneous cloud system

Workflow scheduling has been widely adopted as a deep concern for heterogeneous cloud computing environment and recognized as a well-known NP-complete problem. The scheduling goal of any workflow scheduler is to map the user requests on the available virtual machines (VMs) which are deployed on various cloud servers such that overall processing time and incurred cost is minimized. This paper proposes a delay-based workflow scheduling algorithm for cost optimization in the heterogeneous cloud system. The proposed scheduling heuristic works in two phases. In the first phase, the proposed heuristic introduces a new priority scheme for the tasks. The prioritized tasks are then assigned to the appropriate VMs in the next phase. The algorithm is simulated on various synthetic workflows and two benchmark scientific workflows with different range of tasks. The experimental results of the proposed algorithm supercedes the existing scheduling algorithms in terms of makespan, schedule length ratio (SLR) and cost.