Web service recommendation framework using QoS based discovery and ranking process

Abstract

Locating the desired web service to a client requirement is an onerous task as many web services are ready to satisfy a request. Recommending the pertinent web service and not providing the unwarranted service are the two main issues to be addressed in web service selection process. The limitation of keyword search is that it doesn't articulate the client request accurately. Current search engines and UDDI are based on keyword search, which is inherently impractical for providing the right service to the client. Since the keyword search is unable to grasp the underlying semantics of the searched query, web services discovery based on input and output operation has been introduced. Though the operation based search returns relevant result, still the returned result is not ranked based on performance of web services. So the search mechanism must be incorporated with the QoS requirements will augment to complete the selection process. Therefore, QoS aware web service discovery based on input and output operation has been introduced. In this approach, QoSConsultant acts as a broker between client and service provider. The QoS attributes are normalized, and the match score value is stored in the Service pool. The WSDLParser extracts the input/output operation from the WSDL file and stores it in the service pool. Whenever a search is performed for a given input/output operation, the consultant selects a list of candidate services that are matched with the given request will be provided to the client for setting weights over QoS attributes. The highest degree of the matched web service will be provided to the client for invocation process. The system has been tested with real and synthetic data which shows propitious result.