Discovering Service Similarity by Testing

We describe a comprehensive methodology for discovering service similarity (substitutability) by testing. Our solution does not rely on the service descriptions provided by their authors, it does not assume the existence of ontologies, semantic tagging, or other representations, and it avoids common information retrieval techniques. The only information our method expects is that every service in the repository will be annotated with its inputs and outputs, and that each input and output will be associated with a domain from a published list.