Semantic Service Discovery Ontology based Indexing

A two-step indexing based method which indexes services by their ontologies is proposed for efficient and quick semantic service discovery. In first step, for a given service request, a set of candidate services are chosen from the index by matching the ontologies of the request with the keys of index. In the second step, the request is semantically matched with candidate services ranked by their similarity score. The indexing of services helps in eliminating the irrelevant services of a request. Semantic matching will be performed only to candidate services rather than all available services. From experimentation, it is found that the proposed method quickens service discovery by an average elimination of irrelevancy of 9 measures such as ‘service loading time’ and ‘service and service matching time of the proposed method are significantly reduced to 1 seconds when compared to sequential m matching time of 577.7 seconds and 353 which indexes services by outputs, the proposed work exhibits excellent recall and precision Keywordsindexing based service discovery, indexing by ontologies, service discovery In Service Oriented Computing (SOC), complex business transactions are composition where atomic services are In real time applications, manual discovery is complex due to the existence of huge number of Web. So, services should be automatically discovered and combined. Service description languages such as Service Description Language (WSDL) automate service discovery, services are described using and discovery bring in maximum automation in its accuracy, semantic discovery consumes huge time semantic relations such as equivalent, subsumes, plug Ontology reasoner such as Pellet is used to find the semantic relations between concepts framework, there are two major components namely, matching algorithm discovers matched se transaction involves multiple services to be composed in complex chain become critical needs in service composition based applications introduced either in semantic reasoner or in architecture of a reasoner, the second method is extensively used. Predominantly used techniques for optimization at matcher include clustering [ In this work, a two-step inverted indexing based proposed for efficient and quick semantic candidate services which have ontologies same as that of services are indexed by their ontologies considered as irrelevant to the request semantically matched with candidate services to are ranked by their similarity score. In the existing indexing based service discovery used as keys of index and each key is linked to a list of services that contain that key. This is shown in Fig 1. For June 2012 Semantic Service Discovery Ontology based Indexing School of Computer Science and Engineering Bharathidasan University Tiruchirappalli, TN, 620 024, India , [email protected]