Design And Implementation Of Service Discovery Function Based On Interface Matching In Web Service Search Engine

Over the past years, Internet has developed so much and been very popular. Web services have become an important platform to develop Web applications, as they can support interoperability among applications at heterogeneous platforms. With rapidly developing of Web services, a growing number of Web services are available on the Internet. How to enable users to quickly and accurately find the services they need in the registry having large-scale Web services, has been a problem must be solved in order to continue developing Web services, and also become a research hotspot at both domestic and overseas.At first, the Web service discovery mechanism is based on the UDDI (Universal Description, Discovery and Integration). UDDI is a directory service, which supports keyword-based search method, but it just stays at the syntax level and lacks semantic reasoning support. Then, the researchers have introduced the idea of the Semantic Web, trying to improve the recall rate and the precision of Web service discovery. As Semantic Web services need to change the service implementation support agreement, and a lot of the relative technology is not yet mature, it cannot be used for commercial operation.Apart from keyword-based service discovery mechanism and semantic-based service discovery mechanisms, a new service discovery mechanism was proposed at the first stage of project on the subject of "Converged Network Business System Development" which is relied on by this article. The mechanism first abstracts Web services to interface entity which has input and output parameters, then finds matching Web services by matching the interface information, and is called a interface-based matching service discovery mechanism.Because there is multi-dimensional interface information, we need to integrate a variety of decision-making to match services and calculate the multiple forms of matching. Matching algorithms for each part of interface information have been carried out at the first phase of the project, but its research work needs to be further improved. This article summarized the pattern that can match the number of parameters, parameter order, parameter types, keyword information and semantic information of parameter names. Therefore, the design programs for each matching pattern are the key technologies of service interface matching.For Parameter type matching model design, we take it into account that Web services can support cross-platform data type system and interface parameters actually have many types. Namely, the type of parameters can be built-in data types, and can also be custom type, or they can be simple data types, opposite to complex data types. For the matching of complex types, we introduce a bipartite graph model in graph theory. It is the important research content of this article that how to give a complex type atomic operation by its member data types, and then determine the matching degree between request and target complex parameter types by a bipartite graph matching method.In addition, this article focuses on the design and implementation of service discovery function based on interface matching. As part of the service search engine, the service interface matching subsystem implements this function. This article first introduces the system functional requirements, then tells its overall design, divides the subsystem into functional modules, designs the operating processes of each module, finally verifies its function using test cases.Finally, we summarize the research conclusion has been made, then identify problems, explore available access to studying the problem and point out the next step in the research work.