Supporting molecular modeling workflows within a Grid services cloud

Abstract

Seamless integrated support for scientific workflows accessing HPC applications, deployed on globally distributed computing resources, has become a major challenge in scientific computing. Scientific workflows in the domain of theoretical chemistry are typically long running, deal with huge files, and have a need for dynamic execution control mechanisms. In this paper, we describe a service-oriented approach based on the Vienna Grid Environment (VGE) that tackles these challenges by seamlessly integrating the Ubuntu Cloud infrastructure supporting the scheduling of dynamic and partitioned workflows. The VGE service environment, which enables the provisioning of HPC applications and data sources as Web services, has been enhanced with support for virtualized workflows. The generic scientific workflow infrastructure is utilized in the context of the CPAMMS project, an interdisciplinary research initiative in the area of computational molecular modeling and simulation. A case study implementing a complex scientific workflow for computing photodynamics of biologically relevant molecules, a simulation of the nonadiabatic dynamics of 2,4-pentadieneiminum-cation (Protonated Schiff Base 3, PSB3) solvated in water, is realized via the presented infrastructure. © 2010 Springer-Verlag Berlin Heidelberg.