Automatic Discovery of Web Services Based on Dynamic Black-Box Testing

Most current approaches to Web service discovery exploit description logic-based matchmaking algorithms. Until now, these Semantic Web Service technologies have not substituted for traditional search methods, because higher modeling costs, such as representing service annotation ontologies and domain ontologies, are required. In this paper, we present an automatic testing-based approach to the discovery of Web Services. Through the automatic generation and execution of test cases based on a small amount of semantic information, Web Services are simultaneously evaluated and a decision is made on which Web Services the requestor needs. The experimental results show our technique reduces modeling costs significantly, whilst outperforming existing approaches in terms of precision. Furthermore, reliable QoS attribute values are obtained as by-products.