The Study And Implementation Of Web Services Composition

As a kind of component with excellent capability of well-encapsulation, loose-coupling and cross-platform, Web Service deployed rapidly on the Internet with the development of Internet technology and Internet application. It has already been the focus of current research. There have been more and more Web services on the Internet in recent years. However, single Web Service can not provide complicated functions. Thus, it's necessary to compose existing Web Services to produce new composite Web services which can satisfy the users' requirement. In the process of Web Services composition, one of the key tasks is how to find the related Web Services effectively by using users' input information.Automatic service composition algorithm now can only generate simple Web composition services with only sequence or split logic in the OWL-S specification.It is easy to generate complicated composite Web service by OWL-S specification. OWL-S specification defines OWL-S control constructs:Sequence,Split,Split+Join,Choice,Any-Order,Condition,If-Then-El se, Iterate, Repeat-While and Repeat-Until. Control constructs complicated can be built by these control constructs.Generate fixed composition web service based on the OWL-S specification in this paper. In another word, generate man-made composite services document named templates according to OWL-S specification. Template depository is made from many templates. This paper implements the semi-automatic composition algorithm base on IO match. The nuclear idea that search service in a iterative way from back to front base on the output satisfy requirement firstly,so that find out the templates required by users. Use the model in advance to decompose the search according to the field knowledge on the user request or consult with user as a strategy in order to reduce complexity. In the same time, the templates were stored into template depository using the algorithm.It is a time-consuming task to search the template depository. This paper also proposes a service composition search algorithm with classified cache to respond users' request quickly. The nuclear idea is that saving the templates according to their purposes, and using LRU cache algorithm to improve the probability of hitting the target. Theoretic analysis and experiment results both indicate that the algorithm can improve the efficiency of template search effectively.This paper designs and implements a service composition sub-system architecture which can be expanded based on the Web service composition algorithm and template search algorithm, and also can act as a platform for validating the service composition algorithm.