Developing on-demand secure high-performance computing services for biomedical data analytics.

We propose a technical and process model to support biomedical researchers requiring on-demand high performance computing on potentially sensitive medical datasets. Our approach describes the use of cost-effective, secure and scalable techniques for processing medical information via protected and encrypted computing clusters within a model High Performance Computing (HPC) environment. The process model supports an investigator defined data analytics platform capable of accepting secure data migration from local clinical research data silos into a dedicated analytic environment, and secure environment cleanup upon completion. We define metrics to support the evaluation of this pilot model through performance and stability tests, and describe evaluation of its suitability towards enabling rapid deployment by individual investigators.