Diplomarbeit DIP-2744

Bibliograph.
Daten
Reimann, Peter: Optimization of BPEL/SQL Flows in Federated Database Systems.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Diplomarbeit Nr. 2744 (2008).
191 Seiten, englisch.
CR-Klassif.H.2.4 (Database Management Systems)
H.2.5 (Heterogeneous Databases)
H.4.1 (Office Automation)
K.1 (The Computer Industry)
KeywordsBPEL; BPEL/SQL; business process; federated database system; heterogeneous database systems; optimization; PGM; WebSphere Integration Developer; WebSphere Process Server; WS-BPEL
Kurzfassung

In the Web Services Environment, the standard to model business processes is the Web Services Business Process Execution Language (WS-BPEL or BPEL for short), which has been developed by the Organization for the Advancement of Structured Information Standards (OASIS). In BPEL, it is possible to call Web Services that encapsulate SQL-specific functionality and hide the corresponding data stores. In addition, leading developers of workflow systems (Oracle, Microsoft, IBM) offer the opportunity to tightly integrate SQL-specific functionality into BPEL. By extending BPEL with additional activity types, it is possible to make SQL activities visible in BPEL processes. This facilitates an easier way of modeling data-intensive business processes.

The Institute of Parallel and Distributed Systems at the University of Stuttgart has proposed a heuristic approach to optimize data processing in business processes. This approach is based on rewrite rules which transform data-specific parts of a business process in such a way that the resulting process shows an improved performance and can be executed in a more efficient way. Furthermore, the rewrite rules must not change the semantics of the original process. They are applied to a business process in a multi-stage control strategy. Until now, the approach will only be applicable if some assumptions regarding the involved business process hold. One of these assumptions is that all data management activities of the business process are issued against the same centralized database system. In a practical setting, business processes often access more than one data source, and the involved data sources distinguish themselves by a high degree of heterogeneity. Federated database technology can be used to hide the heterogeneity of multiple data sources and to provide a uniform access to them.

This thesis investigates what changes need to be done to the current, heuristic optimization approach in order to support the optimization of a business process in two new cases regarding the application environment of the process. In the first case, the business process accesses multiple heterogeneous or homogeneous, centralized database systems. In the second one, it accesses one federated database system integrating multiple database systems. Furthermore, new optimization techniques are identified, in particular optimization techniques which are applicable in one or both of the new application environments, and tested for their suitability for the current, heuristic optimization approach. If new optimization techniques are suitable for this approach, according rewrite rules are defined and integrated into the existing set of rewrite rules. The effectiveness of each current and new rewrite rule is tested by measuring the durations of corresponding original and optimized processes.

Volltext und
andere Links
PDF (2296747 Bytes)
Zugriff auf studentische Arbeiten aufgrund vorherrschender Datenschutzbestimmungen nur innerhalb der Fakultät möglich
KontaktSend an e-mail to PeterReimann@web.de
Abteilung(en)Universität Stuttgart, Institut für Parallele und Verteilte Systeme, Anwendersoftware
BetreuerVrhovnik, Marko
Projekt(e)SQL4WL
Eingabedatum18. August 2008
   Publ. Abteilung   Publ. Institut   Publ. Informatik