The increasing popularity of Web Services (WS) has exemplified the need for scalable and robust discovery mechanisms. Although decentralized solutions for discovering WS promise to fulfill these needs, most make limiting assumptions concerning the number of nodes, the topology of the network and having information on the data a-priori (e.g. categorizations or popularity distributions). In addition, most systems are tested via simulations using artificial datasets. In this research we present a lightweight, scalable and robust WS discovery mechanism based on real-time calculation of term popularity.
We believe that this work is unique in this respect. Firstly, we have used a real-world dataset, obtained by processing the information collected by the SeekDa Web Service search engine. This dataset consisted of a set of keywords describing web-services and a set of user queries posted on seekDa.com. Furthermore, we have implemented our algorithm and have deployed it in the context of the OpenKnowledge project. Finally, we have tested it and measured its performance using hundreds of real nodes under heavy load and in the presence of failures.
Anadiotis G, Kotoulas S, Siebes R. 2009. Massively Scalable Web Service Discovery. Submitted for publication at the IEEE 23rd International Conference on Advanced Information Networking and Applications (AINA-09).