Parallel optimization of QoS-aware big service processes with discovery of skyline services

Abstract

With the explosive increase of services in cloud, an important and challenging issue is how to find the best execution plans of big service processes. Although some parallel approaches have been proposed recently, none of them comprehensively considers multiple service processes with various structures, QoS constraints and inter-service correlations, thus cannot precisely evaluate the performance of an execution plan and achieve its satisfaction. In this paper, we present a novel approach for parallel optimization of QoS-aware big service processes with discovery of skyline services. First, a Parallel Discovery Algorithm of Revised Skyline services (PDARS) is proposed, so as to precisely filter out the candidate services for big service processes. Then, a Parallel Meta-heuristic Algorithm considering QoS constraints and Inter-service correlations (PMAQI) is proposed, so as to find the best execution plan of big service processes effectively and efficiently. Finally, experimental results demonstrate that our approach outperforms other methods with higher utility and lower computation time.