In the modern sense of the exploitation of high performance computing of complex special purpose is a complex organizational and technical solution designed to maintain high performance and fault-tolerant information infrastructure in collection, storage, transmission and execution time consuming calculation.
Operating experience the main categories of accidents on high-performance computing complexes of special purpose are failures of engineering equipment and errors of operating personnel. To reduce the failure rate of equipment of the second category of emergencies it is necessary to develop an assistance tool to the operating personnel. Such a tool may be a promising automated control system, including heterogeneous subsystems, covering all aspects of high performance computing systems for military use. The subject of this article is a conceptual study for such a promising automated control system that will provide trouble-free operation of high-performance computing systems for special purposes. Depending on the data that we work with subsystems being considered for the future automated management system can be divided into operational and strategic component. Operational designed for an immediate response with the aim of finding the most rational ways to overcome emergency situations. A strategic component of GRAZING is focused on processing and analysis of large volumes of heterogeneous data collected from the subsystems for reliable forecasting of emergency situations.