SimEA, SyCaTool

The analysis of non-functional properties of single application systems has a long history in the field of Software Engineering. More recently, also techniques and methods have developed that allow the prediction, i.e. the ex-ante analysis, of such properties based on typical early artifacts from the software development process, as e.g. activity and deployment models. Here especially the Palladio Component Model (PCM) developed by Becker et al. [Be09] must be named as a promising model, which is also implemented in a tool, for performing predictions in respect to non-functional properties, as latency or load.

In the field of enterprise architecture (EA) analysis, the situation is different. Here, the topic of analyzing EAs in respect to non-functional properties is relatively new. Even more, the question on how to predict such properties for planned architectures has yet only been explored in a few papers, as e.g. the work of Lankes et al. [La08] or the work of Johnson et al. [Jo06]. Building on this work, Buckl et al. proposed in [Bu09] a stepwise method for conducting ex-ante analyses of EAs based on conceptual models of the corresponding architectures. In this method, the authors further denoted how the expertise of architecture experts can be leveraged in preparing the analyses and establishing an analysis framework.

The importance of ex-ante analyses of EAs is undisputed in literature (see e.g. [Bu09] or [Ia06]). Additionally, industry partners of Systemcartography project research have repeatedly raised interest in quantitative analysis of their EAs, as summarized by Buckl et al. [Bu09a]. This can be exemplified at a partnering enterprise from the banking industry. This enterprise will in the near future undergo extensive changes in its application landscape, as the mainly host-based landscape will be transformed to a distributed one composed of enterprise Java applications. During the preparation of this transformation, major objections on the performance of the future landscape were raised.

In the guided research, we plan to apply the stepwise method for ex-ante analysis outlined by Buckl et al. [Bu09] in the practical environment of the aforementioned enterprise. Thereby, we seek to provide guidance for the evolution of the enterprise’s application landscape towards a distributed business process support infrastructure. For performing the complementing ex-ante analyses, we plan to use the PCM and apply it to plans of the application landscape as part of the EA. We thereby target to establish a generic mechanism for linking structural models of the EA or parts thereof to executable models underlying simulative tools, as the PCM. Based on the results of our analyses, we provide input for EA planning for the banking enterprise.

• [Be09] Becker, S.; Koziolek, H.; Reussner, R.: The Palladio component model for model-driven performance prediction. Journal of Systems and Software, 82:3–22, 2009.
• [Bu09] Buckl, et a.: A pattern-based approach to quantitative enterprise architecture analysis. In 15th Americas Conference on Information Systems (AMCIS), San Francisco, USA, 2009.
• [Bu09a] Buckl, S. et al.: State of the Art in Enterprise Architecture Management 2009. Bericht. Chair for Informatics 19 (sebis), Technische Universität München. Munich, Germany. 2009.
• [Ia06] Iacob, M.E. and Jonkers, H.: Quantitative analysis of enterprise architectures. In D. Konstantas, J.-P. Bourri`eres, M. L´eonard, and N. Boudjlida, editors, Interoperability of Enterprise Software and Applications, pages 239–252, Springer, Geneva, Switzerland, 2006.
• [Jo06] Johnson, P., Nordström, L., Lagerström, R.: Formalizing analysis of enterprise architecture. In: Interoperability for Enterprise Software and Applications Conference, Bordeaux, France, Springer (2006).
• [La08] Lankes, J. and Schweda, C.M.: Using metrics to evaluate failure propagation and failure impacts in application landscapes. In Multikonferenz Wirtschaftsinformatik, GITO-Verlag, Berlin, Germany, 2008.