期刊名称：International Journal of Grid and Distributed Computing
印刷版ISSN：2005-4262
出版年度：2016
卷号：9
期号：4
页码：223-232
DOI：10.14257/ijgdc.2016.9.4.20
出版社：SERSC
摘要：For the large scale services with high-dimensional QoS attributes and distributed environment, traditional service selection approaches are faced with unprecedented challenges in terms of efficiency and performance of QoS. To address these challenges, we propose a three-phase large scale Skyline service selection framework for service composition in clouds. This framework adopts distributed parallel Skyline computation with MapReduce to prune redundant candidate services, and employs parallel multi- objective optimization algorithm based on MapReduce to select Skyline services from the tremendous amount of Skyline services warehouse for composing single service into a set of more powerful Skyline composite services, then applies Top-k query processing technology or multiple attribute decision making support method to select k Skyline composite services from the set of Skyline composite services. Through theoretical analysis, the framework can efficiently solve the service selection problem with large scale services, high-dimensional QoS in cloud computing environment, and quickly generate better composite services with the global optimal QoS.
