Chinese Journal of Computers   Full Text
Title	Research on Service Selection Approach Based on Composite Service Execution Information
Authors	ZHANG Ming-Wei1),2) WEI Wei-Jie1) ZHANG Bin1) ZHANG Xi-Zhe1) ZHU Zhi-Liang2)
Address	1)(School of Information Science and Engineering, Northeastern University, Shenyang 110004) 2)(School of Software, Northeastern University, Shenyang 110004)
Year	2008
Issue	No.8(1398—1411)
Abstract &   Background	Abstract Composite service selection has received increasing attention by the research community in the past few years. Most selection approaches are based on service QoS information which is difficult to acquire accurately, and the corresponding algorithms are complex. In this paper, a composite service execution information based service selection approach is proposed. This approach includes three steps: (1) data producing step: recording the execution information of composite services into a log repository, and extracting related datasets; (2) data mining step: discovering path fork association rules and service execution sequential sequence patterns by data mining; (3) service selecting step: selecting proper services based on the discovered knowledge. In this paper, a composite service system architecture which can record the log easily is presented firstly. Then, a time weighted algorithm mode is proposed to discover the path fork association rules and the service execution sequential sequence patterns. At last, the service selection approach based on composite service execution information is described. Experimental results show that the new selection approach is better than the QoS based approaches on the robustness of the selected composite services. Keywords composite services; service selection; association rule; sequence pattern; data mining Background Composite service selection problem has been extensively studied recently, mostly based on Quality of Service(QoS). The different approaches that have been followed so far in the research of QoS-driven service selection span from the use of QoS ontology, the proposal of ad-hoc methods in some general framework, the use of data mining and so on. Most approaches are complex. Furthermore, current Web service technology can not support QoS or other non-functional aspects of a service, so all the approaches above face the vital problem that how to acquire Web service QoS information. Compared to service QoS information, the composite service execution information is easier to acquire. This paper presents a service selection approach based on composite service execution information, gives a complete solution of the brand new service selection approach, and describes all the related algorithms. The work of this paper is an important part of the authors’ research project, which focuses on key techniques of service management based on usage information, titled "The Acquisition, Reusing and Value Added Propagation of Service Behavior Patterns", including composite service architecture, service log producing, service selection, service composition, and service substitution etc. Now just a few works have been done in this research field at the international level. Mohsen Rouached and Claude Godart proposed a Web service log architecture, but it can’t acquire service process information. Robert Gombotz and Schahram Dustdar presented the problem of Web services interaction mining and workflow mining, but they didn’t give the concrete way to do the mining. George Zheng and Athman Bouguettaya proposed a Web service mining framework, however they didn’t point out how to use it in specific service computing fields. Their work will promote the research in the service mining field, and the work is supported by the National Natural Science Foundation of China (60773218). The research group has developed a platform WAPSbSBP (Web application prototype system based on service behavior patterns) to support the research. Many research papers of the research group are released, some of which have been indexed by EI and ISTP. The work of this paper solves the service selection problem in the project.