Journal of Systems Integration, Vol 5, No 1 (2014)

Font Size:  Small  Medium  Large

Quality of Data Driven Simulation Workflows

Michael Reiter, Uwe Breitenbucher, Oliver Kopp, Dimka Karastoyanova

Abstract


Simulations are long-running computations driven by non-trivial data dependencies. Workflow technology helps to automate these simulations and enable using Quality of Data (QoD) frameworks to determine the goodness of simulation data. However, existing frameworks are specific to scientific domains, individual applications, or proprietary workflow engine extensions. In this paper, we propose a generic approach to use QoD as a uniform means to steer complex interdisciplinary simulations implemented as workflows. The approach enables scientists to specify abstract QoD requirements that are considered to steer the workflow for ensuring a precise final result. To realize these Quality of Data-driven workflows, we present a middleware architecture and a WS-Policy-based language to describe QoD requirements and capabilities. To prove technical feasibility, we present a prototype for controlling and steering simulation workflows and a real world simulation scenario.

Full Text: PDF

DOI: http://dx.doi.org/10.20470/jsi.v5i1.189

ISSN: 1804-2724

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 Czech Republic License.