Towards a Trustable Virtual Organisation

Jun Ho Huh, Andrew Martin
2009 2009 IEEE International Symposium on Parallel and Distributed Processing with Applications  
Introduction A wide range of research is conducted, archived, and reported in the digital economy. Its influence has grown over the years to include various disciplines from science through to finance and industrial engineering. In consequence, different types of distributed systems have been deployed to facilitate the collection, modelling and analysis of the dispersed data; or the sharing of the computational power. A problem arises, however, when the models or data contain sensitive
more » ... on or have commercial value. These then become lucrative targets for attack, and may be copied or modified by adversaries. This is particularly true in many of the scientific disciplines where researcherswho care about the confidentiality of their sensitive data or the integrity of the collected results -are reluctant to exploit the full benefits of distributed computing. Submitting a privileged job to the distributed resources requires prior knowledge of the security standards of all of the target systems -only those running with acceptable security configurations and patch levels should be selected to execute the job. However, this still remains as a 'trust gap' between the job owners' requirements and current technological capabilities, and serves as a barrier to up-take of existing systems. For instance, a Condor system [1] may provide a digital certificate infrastructure for users to identify resource/service providers and vice versa. Without robust mechanisms to safeguard the keys from theft, however, this solution offers only a modest improvement over legacy architectures. Moreover, rogue administrators might replace the compute nodes with malicious ones, tamper with them, or subvert their security configurations to steal sensitive data and/or return fabricated results.
doi:10.1109/ispa.2009.72 dblp:conf/ispa/HuhM09 fatcat:owr7ow6dofbahdg7fibur5hy2m