Research On Matching Metric And Adaptation Of Session-based Web Service

With the rapid scientific development of technologies such as service-oriented and cloud computing,more and more service developers are publishing web services on cloud platforms,resulting in an explosion in the number of services on the web.However,the increasing complexity of individual services with limited functionality and difficulty in meeting user needs has created a need to generate large granular services that are powerful and meet user needs in a combined form for users to use.How to find matching services for service composition from a large number of services,how to find near-matching services with high similarity if no exact match can be found,and how to convert near-matching services into matchable services through the adaptation mechanism are the key issues that need to be addressed in web service composition.To address the above issues,this paper focuses on the service matching determination,matching metric and adaptation of Web service composition.Firstly,the current state of research on Web service matching determination,matching metrics and service adaptation at home and abroad will be briefly described;then,the shortcomings of the existing research will be further analysed;finally,the focus will be on the relevant aspects such as matching determination framework,matching metrics and adapter generation.The main work includes the following aspects:(1)Building a matching decision framework for Web services.Firstly,the SxSTS model,a formal model covering service syntax,semantics and behavioural protocols,is used to fully characterise Web services;then,a dynamic matching framework is constructed which contains three internal variation processing strategies,namely strong,weak and trace strategies,and defines four types of matching concepts,namely full matching,inclusive matching,partial matching and no matching.Finally,the matching determination algorithm is designed to determine the matching type by simulating the existence of deadlock in the SxSTS product,so as to meet the needs of matching determination of Web services in different situations.(2)Propose a matching metric for Web services.For incomplete matching services,firstly,their mismatch situations are analysed to determine the reasons for incomplete matching;then,static matching metrics are carried out,specifically the message interface matching similarity metric in the syntax layer and the ontology concept matching similarity metric in the semantic layer;finally,dynamic matching metrics are carried out using forward and backward propagation to ensure the accuracy of the overall action process matching similarity metrics.(3)Web service adaptation analysis method is proposed.Services with high matching similarity are selected for adaptation.Firstly,the action mapping is constructed using the correspondence between the parameters in the sending action and the receiving action,and the adaptability is analysed based on the mapping results;then,the adaptability matching analysis based on the automatic action mapping is used to analyse the adaptation policy,and the adaptor is designed to automatically construct and generate the adaptability contract;finally,the correctness of the adaptor is verified by checking whether there is any deadlock in the interaction product between the adaptability contract and the STS of the participating matching services.Finally,the correctness of the adapter is verified by checking whether there is any deadlock in the interaction product of the matchability adaptation contract and the participating STS.(4)For the above theoretical framework and method,a system prototype containing SxSTS model automatic generation,Web service matching determination,Web service matching metric and Web service adapter generation modules is developed to verify the effectiveness of the theoretical framework and method through examples.