Medical Image Processing Workflow Support on the EGEE Grid with Taverna

Ketan Maheshwari (1) , Carole Goble (2) , Paolo Missier (2) , Johan Montagnat (1)

1 Laboratoire d'Informatique, Signaux, et Systèmes de Sophia-Antipolis (I3S) / Equipe MODALIS
2 School of Computer Science [Manchester]

Resource-intensive and complex medical imaging ap- plications can benefit from the use of scientific workflow technology for their design, rapid implementation and reuse, but at the same time they require a Grid comput- ing infrastructure to execute efficiently. In this paper we describe a technical architecture that bridges the gap between the Taverna workflow management system and the EGEE grid infrastructure. This is achieved through a novel Taverna gLite activity plugin that makes the in- terface between the multi-threaded, centralized workflow enactor and the massively distributed, batch-oriented grid infrastructure. The plugin significantly increases the performance of Medical Imaging workflows over an equivalent plain Taverna workflow.

Domaines: Calcul parallèle, distribué et partagé [cs.DC]

Ketan Maheshwari, Carole Goble, Paolo Missier, Johan Montagnat. Medical Image Processing Workflow Support on the EGEE Grid with Taverna. 22nd IEEE International Symposium on Computer Based Medical Systems (CBMS'09), Aug 2009, New Mexico, United States. pp.1-7, ⟨10.1109/CBMS.2009.5255332⟩. ⟨hal-00677802⟩