Artikel in Tagungsband INPROC-2010-07

Bibliograph.
Daten
Scheibler, Thorsten; Roller, Dieter; Leymann, Frank: Executing Pipes-and-Filters with Workflows.
In: ICIW 2010.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik.
S. 1-6, englisch.
IEEE Computer Society, Mai 2010.
Artikel in Tagungsband (Konferenz-Beitrag).
CR-Klassif.H.4.1 (Office Automation)
Kurzfassung

In the context of Enterprise Application Integration (EAI) individual tasks have typically been implemented using specialized EAI-vendor technology based on messaging, and quite often customer-specific implementations. Those implementations prominently exploit the Pipes-and-Filters (PaF) architecture. This implementation approach is in conflict with the flow technology that is a cornerstone of the Service-Oriented Architecture (SOA) for solving integration problems. In this paper, we present how this conflict can be resolved. Therefore, we first analyse the differences between PaF architectures and (work-) flow approaches in general, and present the advantages and drawbacks of each solution. Afterwards, we describe how integration solutions modelled based on the PaF architecture can be implemented using flow technology by transforming the appropriate PaF patterns, in particular those used in EAI, into appropriate WS-BPEL constructs. We then present the results of appropriate tests that show that the performance of the corresponding workflows is superior to the mapping of PaF patterns to message flows. We finish off with outlining the additional tangible and non-tangible benefits that the Workflow Management System (WfMS) provides, such as monitoring. In a nut shell, we illustrate that the PaF architecture does not require an own implementation. It is sufficient to have a PaF modelling tool and then convert the appropriate

Volltext und
andere Links
http://www.iaria.org/conferences2010/ICIW10.html
Abteilung(en)Universität Stuttgart, Institut für Architektur von Anwendungssystemen
Eingabedatum3. Februar 2010
   Publ. Institut   Publ. Informatik