Department of Distributed and Multimedia Systems, University of Vienna
In this thesis a thorough description of UN/CEFACT´s Modeling Methodology (UMM) was given. We started with identifying problems and needs in real-world B2B scenarios. Based on these needs we depicted the requirements for a methodology like UMM in order to denote cross-enterprise business processes. Participating in such an inter-organizational process requires a partner to agree on a certain choreography. Furthermore, a participant must provide compliant interfaces to his information systems according to the agreed choreography. However, if each participant describes the same process just from its own view, the resulting process descriptions won´t match. Thus, we require an approach - like UMM - in order to specify process from a global point of view. <br> Moreover, we depicted the need for an adequate tool support in order to endorse modelers in producing UMM compliant models. Afterwards, we presented some technical details in respect to our implementation called UMM Add-In. The UMM Add-In in an extension to the UML tool Enterprise Architect written in .NET. <br> After giving some general information about UMM´s history and the responsible standardization body - the United Nations Centre for Trade Facilitation and Electronic Business (UN/CEFACT) - we introduced worksheets and their role in UMM to capture requirements. In addition, we described the integration of a worksheet editor into the UMM Add-In as well as the benefits thereof. <br> The next chapter comprises a guide supporting the creation of a UMM compliant model based on the know-how gathered by worksheets. The guide explains step by step the modeling of each view in the UMM. Furthermore, the required steps to draw up a UMM model are illustrated using an example about an ordering process.<br> However, giving only a guide for modeling UMM might result in models, which are not compliant to the UMM meta model defined by the UML profile for UMM. This is not a modeler´s failure, but a presumable result when applying a formal notation like UMM without a computational verification. Hence, we illustrated the need for validating UMM models and gave an introduction into the UML profile for UMM. Furthermore, we presented the implementation of a validation engine as part of the UMM Add-In based on the constraints defined in the UML profile for UMM.<br> In a service-oriented architecture, XML-based process descriptions are utilized to configure information systems according to a particular process. In our case, we have already modeled collaborative processes by means of UMM. It follows, that an automatic generation of process descriptions is required in order to support the deployment of B2B information systems. In the UMM Add-In we implemented a transformation engine for generating BPEL compliant artifacts. BPEL seems to be the winner amongst the set of lately emerged choreography languages. <br> Modeling exchanged business information and deriving document schemes thereof is the last building block to gain a complete set of artifacts that together describe a collaborative process. Describing the derivation of schemes as well as a reference implementation thereof according to rules standardized by UN/CEFACT is subject to the following chapter. <br> Finally, in the appendix we included the UML profile for UMM, denoting the UMM meta model. During the writings of our thesis we attended the UN/CEFACT conferences and participated a lot in the development of this profile. <br> However, we think that with our thesis a cornerstone has been laid, which allows the development of additional features. This might include the development of a registry connector in order to support the registration and retrieval of business collaboration models or parts thereof. In addition, features like the generation of BPEL artifacts or document schemes require some improvements in order to be ready for use in production environments. <br> We hope, that the UMM Add-In and this thesis will help to increase the diffusion of UMM within the business process modeling world. Furthermore, our work should be a step forward to achieve the transition from a data-centric business modeling to a process-centric business modeling.
Computational Science and Engineering