Abstract
As the reliability on clouds as an effective platform for scientific workflow applications increase, cloud computing structures are gradually becoming heterogeneous. In service-oriented systems that are heterogeneous, the management of the reliability of resources is a critical issue. Due to this heterogeneity, different models and bandwidths, existing workflows mainly focus on traditional computing environments such as grids. This paper proposes a deadline constrained scheduling strategy for scientific workflows across multiple clouds. To reduce the execution cost and meet deadline at the same time, an adaptive particle swarm optimisation (PSO) technique is proposed. This strategy uses a random single point mutation operator and a two-point crossover operator based on genetic algorithm (GA) technique to optimise computation and data transfer cost. The proposed scheme is evaluated using well-studied scientific workflows. The results obtained from simulations show that the adaptive PSO strategy performs better than existing state-of-the-art scheduling workflow strategies.