Research On Multi-granularity-based Web Service Composition And Collaboration

With the deepening and development of network technology and applications, on the one hand, more and more enterprises and organizations publish their own business function units to the network in the form of Web services, on the other hand, ordinary users put forward higher requests for network services than before. They are no longer satisfied with the simple services, such as weather forecast, flight information search, hotel reservation and so on, provided by the Internet. Users desire high quality services which provide composite functions aiming at their own service requests through the network. It’s a challenge to the current mainstream Web service provider. Apparently, any single provider can’t provide impeccable services for users. In the increasingly competitive market, in order to satisfy the users’composite service requests, the business entities on the same commercial value chain need to form a dynamic enterprise alliance to integrate their own preponderant resources and develop collaborative commerce to aggregate the Web services published by each member of alliance. They form large particle composite services with internal processing logic to dynamically meet the constantly changing users’needs flexibly and quickly. So we research the multi-granularity based Web service polymerization and collaboration in this paper.This paper puts forward a granularity based approach for Web service polymerization and collaboration aiming at real-time, customizability and ambiguity of users’requests. We mainly focus on web service polymerization and collaboration which have large influence on service life cycle.Firstly, this paper first assesses the service quality in service selection which is a basic problem of the service composition process which is based on QoS-granularity and we also give corresponding QoS evaluation method which is based on multi-granularity. We put forward a service quality comprehensive evaluation model. This model comprehensively considers the influence factors of QoS (Quality of Service), including language attributes and interval attributes, in the evaluation process and specially takes user preference and relationship among the different kind services into consideration. In additionally, it gives reasonable definition and executable evaluation calculation formula for each QoS attribute. This model provides foundation for web service composition.Secondly, we put forward a Skyline based multi-granularity QoS service selection method to improve the efficiency of existing QoS-based service selection methods. It selects the Skyline services from candidate service set, and we can prove that the optimal service composition is composed by Skyline services if there is. This method can reduce the number of candidate services to choose appropriate services for service composition effectively and quickly and screen multi-granularity services with the help of Skyline approach.At last, this paper proposes an approach for global optimal service composition based on viterbi algorithm after comprehensively considering the elements of web service composition. Firstly, we build a service composition model presented by a digraph and give the corresponding mapping conversion algorithm. Then, we propose a multi-granularity based approach for the global optimal service automatic composition on the basis of the digraph model. Finally, we prove the correctness and effectiveness of the approach proposed in this paper through comparative experiments and it can be applied to high complexity of web service composition request.We research several key issues in the process of Web service composition and give the corresponding solutions in this paper. It provides a foundation for granularity based web service aggregation and collaboration.