The goal of this research is to realize the vision of a self-organizing linked process for scientific data. This could be an alternative to the current approach to model complex scientific experiments as a graph of dependent tasks. In this approach, processes are used as a method to link scientific data at different stages. Workflow management systems will remain as a planning scheduling engine but will not be forced onto scientists to design experiment. This ecosystem can be thought of as a semantically annotated network where data and processes are interlinked to describe meaningful scientific data transformations.

In practice, this research will assist scientists to identify the data transformation required for his/her experiment and schedule them optimally on appropriate computing resources. This data-centric type of research is becoming more relevant nowadays as e-infrastructures (Grids and Clouds) are more robust allowing applications to reach unprecedented scales in data exploration.

 

For more details about my research interest visit the data-centric research web page of the System and Network Engineering group