Achieving Interoperation of Grid Data Resources via Workflow Level Integration

Abstract

Production Grids are becoming widely utilized by the e-Science community to run computation and data intensive experiments more efficiently. Unfortunately, different production Grid infrastructures are based on different middleware technologies, both for computation and for data access. Although there is significant effort from the Grid community to standardize the underlying middleware, solutions that allow existing nonstandard tools to interoperate are one of the major concerns of Grid users today. This paper describes the generic requirements towards the interoperation of Grid data resources within computational workflows, and suggests integration techniques that allow workflow engines to access various heterogeneous data resources during workflow execution. Reference implementations of these techniques are presented and recommendations on their applicability and suitability are made.

Scientific workflows: Past, present and future

“…Over the last decade Galaxy has been referenced, used, implemented, and or extended in over 4300 publications, including over 3500 journal articles. continued on next page continued from previous page Advances and successes in the past 10 years WFS Advances or major achievements gUse Kacsuk et al [45] and Kiss et al [46] demonstrated, using WS-PGRADE, how scientific workflows could tackle the interoperability of computational and data resources of Grid computing architectures. A coarse-grained workflow interoperability approach that enables embedding non-native workflows into a hosting meta-workflow has been described by Terstysanszky et al [47].…”