Reducing the irrelevant web services and not missing the most relevant web services are the two complimentary issues in web service discovery process. Due to the lack of support from Universal Business Registries (UBR) for complex queries, web service search engine has been introduced to discover the pertinent web services. Traditionally, web services can be discovered by keywords. Since the keywords are unable to bear the underlying the semantics of the search query, operation based discovery has been introduced. Furthermore, clients search interest often changed, we propose a model to discover similar web services based on QoS requirements. To achieve this, find the degree of the functional component that matches with the search query can be computed, and its respective QoS attributes that can be normalized by the proposed algorithm. These two factors can be aggregated to provide overall score value for the web service. For a search request, the list of candidate services will be provided to the client based on the degree of closeness with the search query. The system has been tested with the real and synthetic data, and the test result shows very favorable results.

Key words: web services, QoS attributes, Normalization, web service discovery.