Selecting Skyline Services for QoS-aware Composition by Upgrading MapReduce Paradigm

Jian Wu,Liang Chen,Qi Yu,Li Kuang,Yilun Wang,Zhaohui Wu

DOI: https://doi.org/10.1007/s10586-012-0240-9

2013-01-01

Cluster Computing

Abstract:With the development of web technologies and cloud computing, more and more services which provide similar functionality but differ in QoS are deployed on the Internet via cloud platforms. Recently, skyline analysis is adopted to select candidate services with better QoS to facilitate the process of QoS-aware service composition. However, the fast increasing number of services, multiple QoS attributes to be considered, and dynamic service environment pose a big challenge to skyline service selection. In this paper, we present a parallel skyline service selection method to improve the efficiency by upgrading the MapReduce paradigm. An angle-based dataspace partitioning approach is employed in our MapReduce based skyline service selection. In particular, we explore the dominance power of local skyline services to improve the efficiency of selection, and present two detailed algorithms. To handle the dynamic nature of service environment, we employ Paper-Tape (PT) model which is used to rapidly locate varying services, and present a dynamic skyline service selection algorithm based on PT model. By experimenting over both real and synthetical datasets, we demonstrate the efficiency of our proposed methods.