Applying Provenance to Protect Attribution in Distributed Computational Scientific Experiments

Abstract. The automation of large scale computational scientific experiments can be accomplished with the use of scientific workflow management systems, which allow for the definition of their activities and data dependencies. The man-ual analysis of the data resulting from their execution is burdensome, due to the usually large amounts of information. Provenance systems can be used to support this task since they gather details about the design and execution of these exper-iments. However, provenance information disclosure can also be seen as a threat to correct attribution, if the proper security mechanisms are not in place to protect it. In this article, we address the problem of providing adequate security controls for protecting provenance information taking into account requirements that are specific to e-Science. Kairos, a provenance security architecture, is proposed to protect both prospective and retrospective provenance, in order to reduce the risk of intellectual property disputes in computational scientific experiments.