Y. Luo (Ed.): CDVE 2008, LNCS 5220, pp. 54–57, 2008. © Springer-Verlag Berlin Heidelberg 2008

An Ontology–Based Semantic Cooperation Framework for Business Processes

Yue Ni1,2, Shuangxi Huang1, and Yushun Fan1

1 Department of Automation, Tsinghua University, 100084 Beijing, China 2 Guilin Air Force Academy, 541003 Guilin, China

niy04@mails.tsinghua.edu.cn, {huangsx,fanyus}@tsinghua.edu.cn

Abstract. Nowadays many functions within intra- and inter- enterprises have been encapsulated into Web services to enable platform independent integration and interoperation. To realize the enterprise applications, most Web services are composed as workflows based on Business Process Execute Language (BPEL). However, the absence of semantic information in processes reduces the efficiency of interoperation and cooperation among these Web services from different providers. In this paper, a novel framework to support semantic cooperation of business processes is proposed, which brings semantic information into processes through mapping OWL-S to BPEL and Web Service Definition Language (WSDL) documents, then Web services from various domains and enterprises can understand each other eventually. The framework is being implemented under Project ImportNET.

Keywords: ontology, semantic, cooperation, BPEL.

1 Introduction

Recently, Web services have become the dominating technology to construct business processes. Most Web services come from different enterprises in different domains, which bring the problem of heterogeneities for integrating these business processes. Ontology defines a common vocabulary for stakeholders who need to share information within a domain. It includes machine-interpretable definitions of basic concepts in the domain and relations among them. So ontologies are a suitable “single source of truth” to guide and coordinate the interoperability of business processes in order to disambiguate.

Many studies have been dedicated to enterprises integration and interoperability under Service-Oriented Architecture (SOA). Some of these works approached to develop semantic Web services by which the Web services are annotated based on shared ontologies. LSDIS Lab in University of Georgia discussed one approach that added semantics to WSDL using DAML+OIL ontologies [1]. Martin et al. elaborated OWL-S particularly in [2] and semantics was added to Web services by extending BPEL and WSDL with OWL-S. Aslam and Shen [3, 4] presented a method mapping BPEL to OWL-S and implemented a mapping tool. D.H. Akehurst et al. [5] proposed a framework SiTra for the transformation from OWL-S to BPEL. These work concentrated on adding semantics to BPEL and WSDL which was useful in bringing semantics to processes and Web services but insufficient in reflecting the original extractive semantics from the real world. In this paper, we represented an approach to

An Ontology–Based Semantic Cooperation Framework for Business Processes 55

bridge the real-world semantics to processes semantics directly in order to achieve the semantic cooperation for business processes. We use collaboration ontologies which are generated from reference ontologies as the input of Ontology Generator Tool (OGT) so that business processes and descriptions of Web services could keep original semantics.

2 Semantic Cooperation Framework

In order to realize the semantic interoperation of BPEL-based business processes, a cooperation framework is proposed, as shown in Figure 1.

Fig. 1. BPEL-based semantic cooperation framework

Firstly, reference ontologies are extracted from corresponding domains and then as the input of OGT; the output of OGT are *.bpel and *.wsdl files; then *.bpel is loaded into Workflow Engine when receive the request message from Cross Domain User Interface (CDUI) where the request of domain engineer is transformed to SOAP message, a process will be initiated by the SOAP message; finally, the process invokes and executes corresponding Web services according to the *.wsdl files to fulfill the request and returns the result to CDUI.

OGT is the kernel of our framework. Figure 2 shows the process of the mapping from OWL to BEPL and WSDL, this is the main function of OGT. The reference ontologies are transformed by ontology transformer from domain ontologies which belong to specific domains. In our case the domain ontologies are the mechanical ontology and electronical ontology built by Applied Logic Laboratory in Hungary and Salzburg Research in Austria, and both of them are partners in project ImportNET [6]. Presently, the reference ontology is based on the Extended Description and Situation module of the DOLCE top ontology library [7]. In order to keep the generality, the ontologies in different formats are transformed to a standard form which also uses the DOLCE ontology library, but only the top DOLCE Lite module [6].

56 Y. Ni, S. Huang, and Y. Fan

Fig. 2. Operation mechanism for Ontology Generator Tool

Reference ontology is the standardized reference ontology in the form of *.owl file created by Protégé ontology editor 3.4 beta and generated semi-automatically. Concepts of reference ontology are selected automatically by processing relevant documents or manually. Both deletion and selection are realized by labeling concepts in the reference ontology. The selected concepts are integrated into the complete collaboration ontology automatically. After integration, the rough collaboration ontology will be customized: more ontology items could be added into it from the reference ontology, or ontology items can be deleted. When the collaboration ontology is generated, the concepts, instances, axioms (restrictions) that don’t occur in the reference ontology may be added to the collaboration ontology. Taking collaboration ontology as input, OWL2BPEL and OWL2WSDL Translator output *.bpel and *.wsdl files respectively so that corresponding processes and web services can be deployed on Workflow Engine. Presently the deployment of processes is done manually. We do not illustrate the OWL2BPEL and OWL2WSDL in this paper for the limitation of length.

All processes and Web services involved in this framework are semantic enabled which make interactions clear not only in the definition phase but also in the exec-ution phase. So they can get a common “understanding” no matter which domains the providers belong to.

3 System Implementation

A prototype system has been realized in the project ImportNET to validate our framework. The main functions of the system include: creating reference ontologies based on DOLCE top ontology library [7] and generating collaboration ontologies by Protégé; mapping *.owl file to *.bpel and *.wsdl files automatically; building a BPEL workflow engine based on ActiveBPEL Community Edition Engine [8].

In order to explain the functions of our system, an application scenario is given here. Suppose there are two enterprises M and E from two domains: mechanics and electronics respectively. Mike is a mechanical engineer in enterprise M and Eric is an electronical engineer in enterprise E. They need to cooperate to design a kind of hub. When Mike changed the position of a hole in a circuit board, Eric should know it to make some changes in the design of corresponding circuit. The problem is that they are in different domains and may not understand each other. Our method is as follows, first construct OWL-S based collaboration ontologies according to the reference ontologies from different domains, then transform them to BPEL and WSDL files to achieve a semantic enabled process, and thus Mike and Eric can

An Ontology–Based Semantic Cooperation Framework for Business Processes 57

understand each other and the design process can proceed automatically. The example is quite simple but illustrates the key problem that we want to solve in this paper, the details can be found in the website of ImportNET (http://www.importnet-project.org).

4 Conclusion and Future Work

In this paper we bring ontology to solve the problems of semantic heterogeneities in the cooperation of business processes. A novel framework is proposed to support this ontology-based integration and interoperation. The current method for generating collaboration ontologies needs some developers’ interventions. In the future, we plan to develop some automatic mechanism to improve the method.

Acknowledgments. We wish to thank all partners involved in the project ImportNET, especially Milan Marinov, Dr. Alexander Mahl, Wernher Behrendt, Manuela Plößnig, Diana Bischof, Suan Terge, Miklós Szőts. We gratefully acknowledge funding from European Commission within the Sixth Framework Programme (FP6-033610), National High-Tech R&D Plan of China (2006AA04Z166) and National Natural Science Foundation of China (60504030).

References

1. Sivashanmugam, K., et al.: Adding Semantics to Web Services Standards. In: Proceedings of the 2003 IEEE International Conference on Web Services, Las Vegas, NV, pp. 395–401 (2003)

2. David, M., et al.: Bringing Semantics to Web Services with OWL-S. World Wide Web 10(3), 243–277 (2007)

3. Aslam, M., Auer, S., Shen, J.: From BPEL4WS Process Model to Full OWL-S Ontology. In: The 3rd European Semantic Web Conference, Budva, Mentenegro (2006)

4. Aslam, M.A., Auer, S., Shen, J., Herrmann, M.: Expressing Business Process Models as OWL-S Ontologies. In: Eder, J., Dustdar, S. (eds.) BPM Workshops 2006. LNCS, vol. 4103, pp. 400–415. Springer, Heidelberg (2006)

5. Bordbar, B., Howells, G., Evans, M., Staikopoulos, A.: Model Transformation from OWL-S to BPEL Via SiTra. In: Akehurst, D.H., Vogel, R., Paige, R.F. (eds.) ECMDA-FA. LNCS, vol. 4530, pp. 43–58. Springer, Heidelberg (2007)

6. ImportNET D1.3, D2.2, http://www.importnet-project.org/ 7. DOLCE, http://www.loa-cnr.it/DOLCE.html 8. ActiveBPEL, http://www.activevos.com/community-open-source.php