Towards a Trustable Virtual Organisation

Abstract

In many scientific disciplines, the models, data and methods used to produce results have significant commercial value. Researchers in these sectors are often unwilling to exploit the full potential of grid computing because there remains a `trust gap' between their security requirements and present solutions. We describe two trustable architectures, one applicable for a computational grid and the other for a data grid. Both allow the participants to verify the security configurations of others as well as report their own through a remote configuration management service. The grid jobs are dispatched to only those trustworthy, and guaranteed to run in protected execution environments. Furthermore, our trustworthy analysis server enables statistical analyses to be performed on sensitive raw data --- collected from multiple domains --- without disclosing it to anyone.