TSC – Triple Space Computing D. Fensel, R. Krummenacher, O. Shafiq, E. Ku ¨ hn, J. Riemer, Y. Ding, B. Draxler Triple Space Computing (TSC) has been proposed as communication and coordination paradigm based on the convergence of space- based computing and the Semantic Web. It acts as a global virtual shared space like middleware to enable communication and coordination of semantic data based on the principle of publish and read. This paper presents an overview of the work in progress under Austrian FIT-IT funded TSC project (http:==tsc.deri.at). It presents the evolution of the TSC framework, overall architecture and its usage by Semantic Web Services. Keywords: communication; coordination; middleware; Semantic Web Services; space-based computing Triple Space Computing – neuartiges Kommunikations- und Koordinationsparadigma. Triple Space Computing (TSC) ist ein neuartiges Kommunikations- und Koordinationsparadigma, welches aus einer Kombination von ,,Tuple Spaces‘‘ und Semantic Web-Technologien entstanden ist. Der global zuga ¨ ngliche virtuelle Space stellt eine Middleware zur Verfu ¨ gung, welche es ermo ¨ glicht, semantische Daten via Publizieren und Lesen auszutauschen. In diesem Artikel wird die Entstehung und Entwicklung vom FIT-IT-Projekt TSC (http:==tsc.deri.at) vorgestellt: die Rahmenbedingungen und Datenmodelle, eine Architektur und Anwendungsbeispiele im Zusammenhang mit Semantic Web Services. Schlu ¨ sselwo ¨ rter: Kommunikation; Koordination; Middleware; Semantic Web Services; Space-based Computing Eingegangen am 13. November 2006, angenommen am 23. November 2006 ß Springer-Verlag 2007 1. Introduction Aiming at enhancing the facilities for automated information pro- cessing on the Internet, Tim Berners-Lee (inventor of the World Wide Web and Director of the W3C) brought up the vision of the Semantic Web. Since existing Web technologies around URI, HTTP, and HTML do not support automated processing of Web content, the aim is to develop technologies that allow describing Web content in a structured manner; furthermore, semantically defined meta-data shall help to overcome the problem of heterogeneity within the Internet as an open and distributed system. Ontologies have been identified as the basic building block for the Semantic Web, as they provide machine-processable, semantic terminology definitions. In conjunction with the idea of the Semantic Web, Web Services are proposed as the technology for automated information pro- cessing, thus combining the benefits of the Web with the strength of component-oriented computation. In fact, Web Services promise to allow automated interaction and seamless integration of several entities of the Web, thus are considered as the technology for next generation information systems with special regard to Enterprise Application Integration, B2B technologies, and e-commerce. As initial Web Service technologies around SOAP, WSDL, and UDDI failed to realize the promise of seamless interoperability, the concept of Semantic Web Services has been conceived. By adding semantics to Web Service descriptions, intelligent inference-based mechanisms shall allow automated discovery, composition, and execution of Web Services. Space-based computing has its roots in parallel processing. Linda was developed by David Gelernter in the mid-1980s at Yale Uni- versity. Initially presented as a partial language design (Gelernter, 1985), it was then recognized as a novel communication model on its own and is now referred to as a coordination language for parallel and distributed programming. Coordination provides the infrastruc- ture for establishing communication and synchronization between activities and for spawning new activities. There are many instantia- tions or implementations of the Linda model, embedding Linda in a concrete host language. Examples include C-Linda, Fortran-Linda and Shared-Prolog. Linda allows defining executions of activities or processes orthogonal to the computation language, i.e. Linda does not care about, how processes do the computation, but only how these processes are created. The Linda model is a memory model. The Linda memory is called tuple space and consists of logical tuples. There are two kinds of tuples. Data tuples are passive and contain static data, process tuples or ‘‘live tuples’’ are active and represent processes under execution. Processes exchange data by writing and reading data tuples to and from the tuple space. In 2003 and 2004 there have been discussions and collaborations involving Tim Berners Lee, Dieter Fensel, Eva Kuehn and Frank Leymann on the relationships between the Semantic Web, Web Services and space-based computing. Based on that, Dieter Fensel published a technical report about ‘‘Triple Based Computing’’ presenting the idea of a semantically enabled, space-based com- munication and coordination middleware as an infrastructure for the Semantic Web and Semantic Web Services. These ideas have been adopted for the research project ‘‘Triple Space Computing’’ (TSC) funded by the Forschung, Innovation, Technologie – Infor- mationstechnologie (FIT-IT) research programme in the pro- gramme line of ‘‘semantic systems and services’’. Triple Space Computing inherits the publication-based communication model from the space-based computing paradigm and extends it with originalarbeiten Elektrotechnik & Informationstechnik (2007) 124/1/2: 31–38. DOI 10.1007/s00502-006-0408-1 Fensel, Dieter, Univ.-Prof. Dipl.-Ing. Dr., Krummenacher, Reto, Dipl.-Ing., Shafiq, Omair, Bakk. techn., Digital Enterprise Research Institute (DERI), University of Innsbruck, 6020 Innsbruck, Austria (E-mail: dieter.fensel@deri.org); Ku ¨ hn, Eva, Ao. Univ.-Prof. Dipl.-Ing. Dr., Riemer, Johannes, Dipl.-Ing., Institute of Computer Languages, Department of Computer Science, Vienna University of Technology, 1040 Vienna, Austria; Ding, Ying, Dr., Electronic Web Services (eWS) GmbH, 6020 Innsbruck, Austria; Draxler, Bernd, Dipl.-Ing., Thonhauser Data Engineering (TDE) GmbH, 8700 Leoben, Austria. heft 1–2.2007 | 31 Ja ¨ nner/Februar 2007 | 124. Jahrgang