Dynamic Web Service discovery architecture based on a novel peer based overlay network q S. Sioutas a,, * , E. Sakkopoulos b , Ch. Makris b , B. Vassiliadis c , A. Tsakalidis b , P. Triantafillou b a Department of Informatics, Ionian University, 7 Tsirigoti Square, 49100 Corfu, Greece b Computer Engineering and Informatics Department, University of Patras, Greece c Computer Science, Hellenic Open University, Patras, Greece article info Article history: Received 18 June 2008 Received in revised form 12 November 2008 Accepted 13 November 2008 Available online 3 December 2008 Keywords: Web Services Discovery Peer to Peer overlay networks Databases abstract Service Oriented Computing and its most famous implementation technology Web Services (WS) are becoming an important enabler of networked business models. Discovery mechanisms are a critical fac- tor to the overall utility of Web Services. So far, discovery mechanisms based on the UDDI standard rely on many centralized and area-specific directories, which poses information stress problems such as per- formance bottlenecks and fault tolerance. In this context, decentralized approaches based on Peer to Peer overlay networks have been proposed by many researchers as a solution. In this paper, we propose a new structured P2P overlay network infrastructure designed for Web Services Discovery. We present theoret- ical analysis backed up by experimental results, showing that the proposed solution outperforms popular decentralized infrastructures for web discovery, Chord (and some of its successors), BATON (and it’s suc- cessor) and Skip-Graphs. Ó 2008 Elsevier Inc. All rights reserved. 1. Introduction The continuous expansion of the Internet and its relating tech- nologies has created new marketing opportunities: traditional monolithic approaches in information system design are giving way to Service Oriented Computing (SOC). SOC supports the devel- opment of applications as if they were a connected network of func- tionalities (services) available, in a network-enabled environment, within and across different organizations (Singh and Huhns, 2005). Web Services (WS), ‘‘a vision of loosely coupled interaction be- tween components, programs, and applications” and a major implementation technology for SOC, is becoming a driver for busi- ness integration (Borenstein and Fox, 2004). Nevertheless, as more and more WS become available by many vendors, an old, search engine, problem is reappearing in a new form: searching (discov- ery) mechanisms of WS are not efficient both in response times and quality of results. Currently, the majority of Web Services are developed for internal enterprise use. Web Services will become more widely adopted in the years to come, allowing much broader intra- and inter-enterprise integration. It is anticipated that there will be an increasing requirement for automated service discovery, enabling further Web Services interaction with even less human effort. The ‘information stress’ problem in highly distributed envi- ronments is one of the most interesting and difficult for computer science, and has already attracted significant attention (Schmidt and Parashar, 2004; Yu et al.,2004a,b; Panagis et al., 2008; Diamad- opoulou et al., 2008; Adamopoulou et al., 2007). Web Service registries are helping to narrow down the negoti- ation and searching time needed for service discovery. Their basic concept lies in the matching mechanism of the contractual and technical profile of the query to that of the WS. In order for the searching procedure to be fast and more importantly automatic, information needs to be machine processable. Semantic Web tech- nologies have already provided the first standards and tools to- wards machine to machine interaction; truly dynamic automatic discovery however is not yet in our grasp. WS technology is still categorised among most recent members of the Web Engineering area and it has already attracted both scientific community and vendor attention (Sycara et al., 2003). Promising steps have been made towards making WS a worldwide supported solution for application to application (A2A) interaction. Several recent tech- nologies have been developed to support functionality in terms of service communication (e.g. SOAP), description (e.g. WSDL), and discovery (e.g. UDDI). At present, Web Services are mainly advertised in catalogues which are based on the Universal Description, Discovery and Integration standard (UDDI). UDDI has become the predominant technological environment for WS Discovery. A large number of such centralized registry implementations for intra-enterprise communication, each focusing on registering services of interest 0164-1212/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jss.2008.11.845 q Preliminary version of this paper was presented in Proceedings of IEEE International Conference on Information Technology: Coding and Computing, Next – Generation Web and Grid Systems (IEEE/ITCC 2005), pp. 193–198. * Corresponding author. E-mail addresses: sioutas@ionio.gr (S. Sioutas), sakkopul@ceid.upatras.gr (E. Sakkopoulos), makri@ceid.upatras.gr (Ch. Makris), bb@eap.gr (B. Vassiliadis), tsak@ceid.upatras.gr (A. Tsakalidis), peter@ceid.upatras.gr (P. Triantafillou). The Journal of Systems and Software 82 (2009) 809–824 Contents lists available at ScienceDirect The Journal of Systems and Software journal homepage: www.elsevier.com/locate/jss Π23 1/16