Data and Knowledge Grids represent emerging and attracting application scenarios for Grid Computing, and pose novel and previously unrecognized challenges to the research community. Basically, Data and Knowledge Grids are found on high-performance Grid infrastructures, and add to the latter meaningful data- and knowledge-oriented abstractions and metaphors that perfectly marry with innovative requirements of modern complex Intelligent Information Systems. To this end, Service-oriented Architectures and Paradigms are the most popular for Grids, and on the whole represent an active and widely recognized area of Grid Computing research. In this paper, we introduce the so-called Grid-based RTSOA frameworks, which essentially combine Grid Computing with real-time service management and execution paradigms, and place emphasis for novel research perspectives in data-intensive e-science Grid applications on real-time bound constraints. Grid-based RTSOA frameworks are then specialized to the particular context of Data Transformation services over Grids, which play a relevant role for both Data and Knowledge Grids. Finally, we complete the main contribution of this paper with a rigorous theoretical model for efficiently supporting Grid-based RTSOA frameworks, with particular emphasis on the context of Data Transformation services over Grids, along with its comprehensive experimental assessment and analysis.
