HPC Blockchain

Science and research funded and performed by the Department of Energy relies significantly on the modeling and simulations that High Performance Computing systems provide. Research performed on these systems can include energy, climate modeling, nuclear weapons and other government funded research by other agencies. Research missions go beyond local agency use and HPC platforms within Department of Energy and other agencies are national assets for use in any United States funded research. Much of this research directly supports the United States security, economic, energy and health/bioscience postures. The Nation’s policy makers and policy decisions depend on the research performed on these High Performance Computing systems. Compromise of these systems could significantly undermine our Nation’s security and safety. To address the need of adding integrity and data provenance to large computational tasks, we propose the adaptation of blockchain technology for verifiably monitoring and logging processes running within a high performance computer. In Phase 1, we will adapt blockchain technology and characterize the metrics that can be monitored and logged and the performance of different levels of process monitoring. Software developed will allow us to distribute a ledger that contains a timestamp when data is received, a list of operations performed on that data, culminating in an accurate history of how final results were constructed from a starting piece of data. It is anticipated that these class of techniques are expandable to other different distributed computation platforms and applications. This will directly benefit DoE and other agency supercomputing and HPC clusters as well as anyone else intent on using large-scale parallel computing to perform any sort of analytics.