Figure 10: Screenshot of the Interface for Semi-automatic Mapping

Semantic discovery of services is done with the help of the mappings that were recorded in UDDI that is mentioned in sections 4.1 and 4.2. The semantic matching algorithm we have implemented is a simple algorithm to perform semantic discovery. The desired properties of the required Web service can be described using a service template (ST). A service template is created by specifying the inputs and outputs using concepts from the domain specific ontology. Matching of ST with the registered services is then carried out using the categorization details mentioned in 4.1. The results are sent back to the Client Peer.

ebXML version 3.0 which is not yet officially released by OASIS committee which is expected⁶ to discusses the distributed registries model. This registry information model of the version [30] discusses supporting cooperating registries and registry federations. Co-operating registries imply that they are associated with each other, meaning that registry object reference can be across registries. “A registry federation is a group of registries that have voluntarily agreed to form a loosely coupled union. Such a federation may be based on common business interests and specialties that the registries may share. Registry federations appear as a single logical registry, to registry clients” [30]. The objective of this initiative shares some of the objectives of our work. MWSDI supports creating registry federation by grouping registries that are mapped to the same node in Registries Ontology. The registry federations discussed in [30] are based on the P2P model. In MWSDI too, the registries are considered as peers. The federated queries discussed in [30] can be executed in MWSDI using the registries ontology. Though this work seems to be closely associated with our work, it focuses mainly on the registry information model and discusses issues like object replication, object relocation and Lifecycle management for forming registry federation. In comparison, our work is not on the data structures of registries but focuses on building a scalable environment for publication and discovery across multiple registries. Our work also suggests protocols for Peer initiation and Client Interaction. In addition our work uses the Registries Ontology to maintain a global view of the registries, associated domains and uses this information during Web service publication and discovery.

Current research in Web services focuses on semantic Web services. Adding semantics to resources like Web services makes them machine processable [12]. The architecture that has been proposed in [31] discusses using semantics at different levels of Web services stack. They discuss having ontology servers and associated repositories to maintain domain concepts as ontologies. In MWSDI, we provide similar functionalities using Operator Peers that maintain Web service registries and provide domain specific ontologies. Since domain specific ontologies provide a better conceptualization of a domain than general purpose ontologies the publication and discovery can be made more meaningful using them.

DAML-S ontology was created to enable the semantic description of Web services [32]. Recent work from DAML-S group [33] proposes using WSDL in addition to DAML-S description to completely describe a Web service In our work, we have annotated WSDL by associating its input and output types to domain specific ontologies As DAML-S is yet to get industry wide acceptance, we chose to use WSDL to add semantics to it. We have used UDDI structures to store the mappings of input and output types in WSDL files to domain specific ontologies. We have adopted this approach similar to the one suggested by [20]. This related work adds semantic matching capability to UDDI, by translating DAML-S representation of a service to UDDI representation so that it can be translated back to DAML-S representation for semantic matching of service specification. MWSDI provides similar functionality using WSDL, which we believe makes it more easily adaptable to the approach and standards the industry has chosen to accept. A detailed description of additional tags and annotations for adding semantics is provided in [25]. As the work to semi-automatically annotate a WSDL with operation-ontology mapping, preconditions and effects is underway, the publication of services discussed in section 4.1 and 4.2 and the discovery of services discussed in section 4.3 do not include operation-ontology mapping, preconditions and effects.

Using semantic metadata leads to the issues of scalable architectures for sharing, maintaining and distributing it. Peer-to-peer networks seem to provide an ideal environment for such systems. Peer-to-peer and Semantic Web issues are discussed in [34] [35]. [36] discusses using peer-to-peer, Semantic Web and Web services as enabling technologies to create a semantic driven service oriented architecture. Our work encompasses contributions from all these areas and provides peer-to-peer environment for Semantic Web service discovery and publication. Semantic gossiping [37] presents an architecture where mappings between schemas are used as a basis for query propagation. Its uses a bottom up approach for semantic agreement in a peer-to-peer environment, where there is no global ontology. However, while MWSDI allows registries to maintain their own domain specific ontologies or schemas, it uses a global ontology to maintain a relationship between registries. Since the premise of our work is to maintain relationships between registries, using Registries Ontology is critical. For semantic agreement, MWSDI allows registry providers to update the Registries Ontology to either relate their registries to existing concepts or to create their own concepts.
6. Conclusion and Future Work

We present techniques and a prototype implementation of MWSDI. Our approach involves creating an infrastructure of registries for semantic publication and discovery of Web services. The primary motivation of our work is the expected growth in the number of registries and the lack of semantics in Web service representation. Our system provides a scalable architecture to access such registries. In addition, we provide semantic publication and discovery capabilities by using a domain specific ontology for each registry. We have presented two algorithms for semantic publication and discovery using WSDL descriptions. Both these algorithms map inputs and outputs of Web services to ontological concepts. Subsequently, searching can be carried out using templates constructed using ontological concepts.

In our approach, we treat a Web service as a black box having a set of inputs and a set of outputs. Annotating these inputs and outputs gives us a significant improvement in discovery and is better than the current approach used by UDDI. However each WSDL description may have a number of operations having different functionalities. Each operation would have its own set of inputs and outputs. For example, the same Web service may have operations for both selling and buying books. We believe our searching algorithms can be significantly improved by two techniques. Firstly the operations themselves should be mapped to concepts in the domain specific ontology which depict functionality. Secondly all inputs and outputs in the WSDL description should not only be mapped to concepts in the domain specific ontology but also grouped according to operations. The domain specific ontologies would have to be modified to maintain concepts that depict functionality along with the already existing input and output concepts. We are currently working on implementing this algorithm.
A significant part of this paper discusses implementation and architecture of the peer-to-peer network used by MWSDI. We discuss how using a peer-to-peer network gives us the scalability and flexibility required for creating an infrastructure for diverse Web service registries. We have tested our work with UDDI registry implementation provided in JWSDP [38]. However, this idea is applicable to any UDDI registry implementation and other type of Web services registries. Issues not covered in this paper that are planned as future enhancements are:

• 
  Redistribution of service publication among registries
  
• 
  Exchange of semantics between registries
  
• 
  Full query support using all kinds of relationships among registries
  
• 
  Adding reliability for the Gateway Peer by replication
  
• 
  Automation registry selection in a domain using techniques for searching relevant
  ontologies discussed in [35]
  
• 
  Study on performance and reliability of the P2P network and implementing
  security measures
  

From the business perspective MWSDI is all about grouping services and distributing them in different registries based on domain specialty, for locating the right services easily. On the other hand, from the technical perspective, MWSDI provides a scalable infrastructure for accessing multiple registries and semantic enhancements to the current service discovery mechanism. We believe that to develop processes in the current network economy [39], architectures like MWSDI will drive the evolution of businesses interactions using Web services. This infrastructure will also help Web services in changing the focus from static to more dynamic business settings.

[1] Universal Description, Discovery and Integration: UDDI Technical White paper. 2000. http://www.uddi.org/pubs/Iru_UDDI_Technical_White_Paper.pdf

[2] E. Box et al., “Simple Object Access Protocol (SOAP) 1.1", May 2000. Available

at http://www.w3.org/TR/SOAP .

[3] E. Christensen et al., “Web Services Description Language (WSDL) 1.1", W3C Note, March 2001, Available at http://www.w3.org/TR/wsdl

[4] http://ws.apache.org/axis/

[5] The Evolution of UDDI, UDDI.org White Paper, 2002 http://www.uddi.org/pubs/the_evolution_of_uddi_20020719.pdf

[6] L. Gallagher, L. Carnahan. A General Purpose Registry/Repository Information Model, Information Technology Laboratory National Institute of Standards and Technology. 2nd Draft - 23 October 2000.

[7] S. Macroibeaird et al. Using UDDI to find ebXML Regs/Reps, April 2001.

http://lists.ebxml.org/archives/ebxml-regrep/200104/pdf00002.pdf

[8] T. R. Gruber, “A Translation Approach to Portable Ontology Specifications.” Knowledge Acquisition, 5(2), 199-220, 1993.

[9] F. Curbera, D. Ehnebuske, D. Rogers, Using WSDL in a UDDI Registry, Version 1.07, UDDI Best Practice, May 21, 2002. http://www.uddi.org/pubs/wsdlbestpractices-V1.07-Open-20020521.pdf

[10] R. Schollmeier, “A Definition of Peer-to-Peer Networking for the Classification of Peer-to-Peer Architectures and Applications.” Proceedings of the First International Conference on Peer-to-Peer Computing, 2001.

[11] F. Curbera, W. Nagy and S. Weerawarana, “Web Services: Why and How.” Workshop on Object-Oriented Web Services – OOPSLA 2001, Tampa, Florida, USA, October 2001.

[12] T. Berners-Lee, J. Hendler and O. Lassila. “The semantic web.” Scientific American, 284(5):34–43, 2001.

[13] T. Gruber. The role of common ontology in achieving sharable, reusable knowledge bases. In J. A. Allen, R. Fikes, and E. Sandewall, editors, , San Mateo, CA, 1991. Morgan Kaufman

[14] V. Kashyap and A. Sheth. Semantics-based Information Brokering. In Proceedings of the Third International Conference on Information and Knowledge Management (CIKM), November 1994.

[15] A. Sheth. Changing focus on interoperability in information systems: From system, syntax, structure to semantics. In M. Goodchild, M. Egenhofer, R. Fegeas, and C. Kottman, editors, Interoperating Geographic Information Systems. Kluwer Publishers, 1998.

[16] A. Sheth, R. Meersman, Amicalola Report: Database and Information System Research Challenges and Opportunities in Semantic Web and Enterprises, SIGMOD Record 31 (4). December 2002.

[17] M. Dumas, J. O'Sullivan, M. Heravizadeh, D. Edmond, A. Hofstede, Towards a semantic framework for service description. In Proc. of the 9th Int. Conf. on Database Semantics, Hong-Kong, April 2001. Kluwer Academic Publishers.

[18] W. Grosso, H. Eriksson, R. Fergerson, J. Gennari, S. Tu, M. Musen., Knowledge Modeling at the Millennium (The Design and Evolution of Protege-2000). 1999.

[19] Core Component Dictionary, ebXML Core Components, May 2001. http://www.ebxml.org/specs/ccDICT.pdf

[21] F. Gandon, Ontology Engineering: A Survey and a Return on Experience, Institut de Recherche en Informatique et Automatique (INRIA), March 2002, ISSN 0249-6399.

ftp://ftp-sop.inria.fr/acacia/fgandon/FabienGandon_RR4396.pdf

[22] M. Denny, Ontology Building: A Survey of Editing Tools, November 06, 2002

http://www.xml.com/pub/a/2002/11/06/ontologies.html

[23] B. Benatallah, M. Dumas, M-C. Fauvet, F. Rabhi, Towards Patterns of Web Services Composition. In S. Gorlatch and F. Rabhi (Eds): "Patterns and Skeletons for Parallel and Distributed Computing". Springer Verlag (UK), 2002.

[24] L. Gong, "JXTA: A Network Programming Environment". IEEE Internet Computing, (5)3:88--95, May/June 2001.

[25] K. Sivashanmugam, K. Verma, A. Sheth, J. Miller, Adding Semantics to Web Services Standards, Proceedings of The International Conference on Web Services, pages 395-401, 2003.

[26] V. Kashyap, A. Sheth, Semantic and Schematic Similarities between Database Objects: A Context-based Approach, VLDB Journal, 5(4), 1996.

[27] J. Madhavan, P. A. Bernstein, and E. Rahm, Generic schema matching with cupid. In Proceedings of the 27th International Conferences on Very Large Databases, pages 49-58, 2001.

[28] J. Cardoso, A. Sheth, (2002) 'Semantic e-Workflow Composition', Technical Report, LSDIS Lab, Computer Science, University of Georgia, July 2002.

[29] A. Patil, S. Oundhakar, and A. Sheth, Semantic Annotation of Web Services, Technical Report, LSDIS Lab, Department of Computer Science, University of Georgia, March 2003.

[31] C. Bussler, D. Fensel, A. Maedche, A Conceptual Architecture for Semantic Web Enabled Web Services, SIGMOD Record 31(4), pages 24-29, 2002.

[32] A. Ankolekar et al., 'DAML-S: Web Service Description for the Semantic Web', in International Semantic Web Conference, Sardinia, Italy, pp. 348-363, 2002.

[33] DAML-S: Semantic Markup for Web Services, The DAML Services Coalition, http://www.daml.org/services/daml-s/0.7/daml-s.pdf

[34] M. Schlosser et al, A Scalable and Ontology-Based P2P Infrastructure for Semantic Web Services, (2002).

[35] M. Arumugam, A. Sheth, and I. B. Arpinar, Towards Peer-to-Peer Semantic Web: A Distributed Environment for Sharing Semantic Knowledge on the Web (2001).

[36] A.Maedche and S.Staab, Services on the Move - Towards P2P-Enabled Semantic Web Services, Proceedings of the Tenth International Conference on Information Technology and Travel & Tourism, ENTER 2003, Helsinki 2003/01/31.

[37] K. Aberer, P. Cudré-Mauroux, M. Hauswirth A Framework for Semantic Gossiping, SIGMOD Record, 31(4), 2002.

[38] Java Web Services Developer Pack, http://java.sun.com/webservices/webservicespack.html

[39] A. Sheth, W. Aalst, I. B. Arpinar, Processes Driving the Networked Economy, IEEE Concurrency, pp. 18-31, July-September 1999 (Vol. 7, No. 3).