Data and Activity Representation for Grid Computing

Computational grids are becoming increasingly popular as an infrastructure for computa- tional science research. The demand for high-level tools and problem solving environments has prompted active research in Grid Computing Environments (GCEs). Many GCEs have been one-o development eorts. More recently, there have been many eorts to dene component ar- chitectures for constructing important pieces of a GCE. This paper examines another approach, based on a `data-centric' framework for building powerful, context-aware GCEs spanning mul- tiple layers of abstraction. We describe a scheme for representing data and activities in a GCE and outline various tools under development which use this representation.