Towards Workflow Allocation using Dragonfly Algorithm in Cloud Computing Environment

The cloud computing platform provides clients with on-demand access to computational resources on pay you go basis for workflow execution to ensure the quality of service (QoS) parameters like cost and execution time, energy, and resource utilization. This paper proposes a metaheuristic algorithm, Dragonfly algorithm (DA) to minimize the makespan of workflow tasks from cloud users submitted for execution. DA utilizes static and dynamic swarming behaviors to explore and exploit in navigation, search the foods, and avoid enemies. It has various phases: separation, alignment, cohesion, attraction, and distraction. The workflow allocator is implemented in MATLAB for the performance analysis. The experimental findings demonstrate that the Dragonfly algorithm (DA) significantly performs better than its peer in terms of fitness value (makespan) and convergence rate under study.