ScholarWorks@숙명여자대학교

초록

In the advent of high performeding and sophisticated devices, explosion of large-scale data and dynamic variation of traffic change are intensifying. Cloud computing enables software-based resource management considering diversity of data services over on-demand computing infrastructure. As real-time analysis of large-scale data and forecasting of data-intensive HPC application are required, software definition technology should be applied to temporal data processing, analysis, and result verification. In this paper, we have investigated trends and software definition technologies of ScienceDMZ, which provides high speed data transmission environment for scientific big data. In order to solve the problem of real-time processing and result verification, we propose a temporal application and integration of data storage and calculation experiment for high performance research with time limit through weather forecast infrastructure. It is also important to identify resource use case patterns based on the understanding of application programs and assign them appropriately. This paper analyzes the runtime characteristics of input data by selecting WRF model as a data - intensive HPC application. We defined the execution structure of the HPC application that allocates the data in a timely manner through the analysis results and raised the importance.

키워드