Design of business choreographies from global and local view
Design of business choreographies from global and local view
Disciplines
Computer Sciences (50%); Economics (50%)
Keywords
-
E-Commerce,
Choreography,
Conceptual Modeling,
Inter-organizational Business Process,
XML,
Interoperability
The proposed project focuses on the support of Business-to-Business e-Commerce. It continues the work done in my Ph.D. Thesis on the impact of the business context on business collaboration models, choreography languages, and business documents. This thesis extended the concepts of the modeling methodology of the United Nation`s Centre for Trade Facilitation and Electronic Business (UN/CEFACT) called UMM. In the meanwhile the proposed extensions have been included into the newest version of UMM of which I am a co-editor. In our research group we have developed a plug-in to a commercial UML tool that fully supports the UML profile defined by UMM. This tool is recommended in the Australian e-government framework GovDex. The UMM is used to model an inter-organizational business process concentrating on the flow of interactions between collaborating business partners and not on their private processes. The resulting choreography is described from a global view. This means that a single model describes the interactions of all parties. This is appropriate to develop so-called industry reference models or standard models. However, an enterprise usually prefers describing the flow of process interactions with their partners from its local view. This means that each partner in a business collaboration has its own model that differs from the partners` ones, but that must be complementary to the partners` ones. A simple example of complementary flows is if one partner performs an invoke activity the other one must perform a receive activity. If each partner develops its local choreography in isolation it is very unlikely that partners will interoperate or in other words they will not be able to do business electronically. Since this is the case in most current approaches, we do not see process interoperability crossing the borders of enterprises. An alternative approach is considering the specifications of a public choreography when constructing the local choreography. The project proposal is based on this paradigm. It focuses on the analysis and design in a development process for inter-organizational business systems. In the project we use UML to denote the analysis and design artefacts. The definition of a local choreography must handle interactions with multiple parties. The global choreography of UMM models defines binary collaborations, i.e. between two partners only. Consequently, a local choreography will span over multiple global choreographies. Each interaction of a local choreography must consider the targets of a single interaction in the global choreography. Furthermore, the local choreography must keep the flow - as defined in the global choreography - for each binary partnership that is part of the multiparty process. There are approaches to automatically transform global choreographies to local ones. However, they are not appropriate for two reasons. Firstly, the UMM activities are less granular than what is needed in the local choreography and secondly the number of parties in the UMM models which is exactly two, does not necessarily match the number of parties in the local choreography. The goal of the project is the support of the analysis and design of local choreographies. It will deliver a UML profile for modeling local choreographies. This profile and its constraints are attuned to the ones of UMM in order to automatically check compliance of local and global choreographies. It is envisioned to collect appropriate global choreographies and parts thereof, to semi-automatically transform these into local representation and to merge them consistently into multi-party local choreographies. This also requires business-context specific registration mechanisms to store and access artefacts of global choreographies. Of course, the software development process does not stop at the UML models. Hence, another aspect of the project is the transformation of the models to XML-based choreography languages which are used in service oriented solutions by applications in order to monitor and/or execute the process.
- Universität Wien - 10%
- University of Technology Sydney - 100%