Declarative techniques for model-driven business process integration

  • Authors:

  • J. Koehler;R. Hauser;S. Sendall;M. Wahler

  • Affiliations:

  • IBM Research Division, Zurich Research Laboratory, Säumerstrasse 4, 8803 Rüschlikon, Switzerland;IBM Research Division, Zurich Research Laboratory, Säumerstrasse 4, 8803 Rüschlikon, Switzerland;IBM Research Division, Zurich Research Laboratory, Säumerstrasse 4, 8803 Rüschlikon, Switzerland;IBM Research Division, Zurich Research Laboratory, Säumerstrasse 4, 8803 Rüschlikon, Switzerland

  • Venue:
  • IBM Systems Journal
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Business process integration and automation are among the most significant factors driving the information technology industry today. In addressing the manifold technology challenges of integration and automation, new standardization efforts aim at improving the interoperability of businesses by moving toward a declarative specification of business processes, that is, one which describes what a business process does and not how it is implemented. At the same time, Model Driven Architecture® focuses on improving the software-engineering methods with which business process solutions are implemented by separating the business or application logic from the underlying platform technology and representing this logic with precise semantic models. In this paper, we present an approach to the model-driven generation of programs in the Business Process Execution Language for Web Services (BPEL4WS), which transforms a graphically represented control-flow model into executable code by using techniques that originated in compiler theory. We discuss the underlying algorithms as well as general questions concerning the representation and analysis of model transformations. We study a declarative representation of transformation rules, where preconditions and postconditions are represented in the Object Constraint Language. By adopting a declarative approach, we pave the way for future automatic consistency checking of transformation rules and bidirectional reconciliation of evolving models.