Posted on by : admin Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

In this section we describes how the Cross Domain Ontology [5] can be used to identify URI. Cross Domain Ontology [5] can be used as a linked data for annotating the web services. The Enhanced version of WSDL[3] with Annotated web services will be deployed as a SAWSDL[2]. In short, input to the process is WSDL [3] and output is SAWSDL [2]. This can be accomplishing with three steps:

1. Prepare WSDL mapping to Ontology.

2. Prepare tokens to annotate.

3. Create a concept from Ontology based on tokens.

A. WSDL to Ontology Mapping Model

WSDL [3] file when submitted will be parsed and elements are divided in to three parts. As only abstract definition of the web service need to be used in composition and discovery it is not required to add annotation to other element. So here we only consider following elements for annotation.

  • Service: Service is the entry point while reading WSDL [3]. It contains information like service name, Service location, Protocol used for executing the service etc. It also contains one or more ports that abstractly describe the operations. Service element of WSDL can be mapped to Class in Object oriented Programming language.
  • Operations: They are abstractly described in the PortTypes. They are actual interfaces to the web service functionality. Operations are the very important elements to be identified during composition process. We can say that all the functionality that the service can provide, listed by operations. They are mapped to the functions in the Programming Languages
  • Inputs and Outputs: This is elements inside the Operations. This can be mapped to the function arguments in programming languages. This is at the leaf of the web service tree. Inputs specify what the operation will take as input for executing Operation. But unlike in programming languages return type of the Operation is specified using Output, it means what web service operation will provide is specified using Output.

Here when we use cross-domain ontology [5], Concepts of different domains to be classified using class and sub classes. Domain of the ontology is being at the root level, contains URI for the instances that uniquely identify the entity or concept. We can use the three level tree of Ontology to annotate the three level tree of WSDL created in previous step for WSDL [3] document as follows.

  • Domain of Ontology: Domain of ontology will specify the knowledge area of ontology while searching for the concept. This can be used to annotate the service level of the WSDL[3] i.e. it will specify URI for service name. It means it stands at the root level of the annotation tree.
  • OWL Class: After classifying service in to specific domain, we can use concept of that specific domain for creating annotation of operation in web service. Each domain service ontology contains categorization of instances in the classes. They can be further divided in to sub class, but for our purpose we also consider them as a class. Ontological inferences can be derived from sub classing the concepts. This stands at second level of the ontology tree to annotate the WSDL [3].
  • OWL Instances: They are the leaf elements of the ontology tree used for annotating the WSDL[3]. Instances are like URI with pattern like Also this URIs is interlinked with each other by RDF triplets. This Ontology Instance URI can be used for annotating the Leaf level of the WSDL [3] tree i.e. Inputs and Outputs.

B. Process of creating Annotations

Annotation basically uses concept of stemming of WSDL element and searching for reference in ontology. So Annotation Process can be divided in three parts.

Extract Meaningful Word:

  • Based on observation of difference web services available on the web, it is derived that most of the web services are having common structure of the name of each element. e.g. currencyService has Operation named getExchange getExchange has input named code and date getExchange has Output rate This structuring can be exploited to extract meaning full word from element name as shown in Algorithm 1.
  • Retrieve Ontology Concept: Now time to retrieve the concepts from ontology for annotation. Jena is used to search ontology for the specific concept.
  • Create ModelReference Attribute: Using the matching concept from the Ontology annotation attribute can be created like SAWSDL:ModelReference=”URI”. This can be added to the WSDL[3] for and WSDL[3] is returned as SAWSDL file. EasySAWSDL and SAWSDL4J is being used for creating new SAWSDL[2] file.