Efficiently supporting secure and reliable collaboration in scientific workflows

Recently, workflow technologies have been increasingly used in scientific communities. Scientists carry out research by employing scientific workflows to automate computing steps, analyze large data sets and integrate distributed computing processes. This is a challenging task because of insecure procedures in a distributed environment. In this paper, we present an access control framework and models for supporting secure and reliable collaboration. The proposed approaches combine control flows and data flow models to describe scientific workflows, and extend the atomicity sphere concept by considering two levels of atomicity abstraction at the level of process as well as at the level of data, in order to maintain the process consistency and the data consistency in the presence of failures. We also present a case study in a scientific research scenario to show the effectiveness of our approaches.