ITER IMAS: Turning Exabytes of fusion data into high value information
ITER (Latin for "The Way"), currently under construction in the south of France, is a large-scale scientific experiment intended to prove the viability of “fusion” as an energy source. In an unprecedented international effort, seven partners—China, the European Union, India, Japan, Korea, Russia and the United States, have pooled their financial and scientific resources to build the biggest fusion reactor in history. ITER will resolve critical scientific and technical issues in order to pave the way to the next step: a demonstration fusion power plant. ITER is one of the most complex scientific and engineering projects in the world today. The complexity of the ITER design has already pushed a whole range of leading-edge technologies to new levels of performance. However, further science and technology is needed to bridge the gap to the commercialization of fusion energy.
The Integrated Modelling and Analysis Suite (IMAS) is a framework to support all the physics modelling needs of ITER. It consists of a unifying data model, an execution framework, a set of HPC codes (parallelized using MPI, OpenMP and CUDA), a set of integrated workflows, and an I/O infrastructure. It will support ITER plasma operations and tokamak physics research through a plethora of activities: modelling confinement, stability, heating, fast ion and burning plasma physics, turbulence, interpreting diagnostic measurements, simulation, plant-protection, validation, verification and big-data visualisation. ITER IMAS workflows are designed to interact with data from external heterogeneous sources (e.g. from ITER diagnostic measurement data, from de-featured CAD models and meshes, with data from other tokamaks around the world etc.) and between the work-flow agents themselves. ITER data acquisition systems are rated at 50GB/s, with plasma durations targeting 400s and typically 20 discharges per day, resulting in a projected 0.4PB of data per day and placing ITER firmly in the Exascale class of next generation data-centric science experiments.
The fusion data model is already object (rather than file) based and the I/O tool-kit within IMAS (the IDAM Integrated Data Access Management platform developed at CCFE) is by design storage format agnostic. IMAS and IDAM therefore offer an ideal test-case for SAGE Object storage server capability. In addition, the processing capabilities of SAGE will be explored by embedding IMAS workflows directly within the storage architecture, to further minimize data movement and provide a direct coupling between the SAGE I/O API and the IMAS workflow agents. In summary, the SAGE platform offers an exciting route by which the Exabytes of data produced by the diagnostic systems on ITER, together with Exascale HPC simulations can be stored, mined and assimilated, turning huge quantities of data into invaluable information.
This video shows the visible light emitted from a typical tokamak plasma – the bright bands of light which look very similar to solar prominences are in fact due to turbulent activity in the relatively cold plasma edge – ions and electrons stream along the field lines and impact the plasma facing components. Understanding the physics of the plasma “exhaust” is a high priority part of the international fusion program. IMAS and SAGE will be used to assimilate huge amounts of diagnostic data and make tractable the Exascale MHD and turbulence simulations that will be instrumental in learning how to control the hot core and interface between the plasma and first wall.
Visit ITER website for more information.
The ITER tokamak is shown below – the outer surface of the plasma is here rendered pink and a standard human man is shown standing within the cryostat bottom left.