The Architectural Design of Storage System for Power Data Management

As interest in the key technologies of the 4th industrial revolution such as artificial intelligence, big data, and the Internet of things has increased, we are developing systems that incorporate ICT technology in various platforms. In particular, the demand for ICT technology in the energy field is increasing. In the power field, it is used to analyze the large-scale power data that is continuously generated and to build forecasting and decision making system for demand. However, in order to analyze such a big data, a machine learning analysis is performed on data generated in a system such as a power equipment in real time, and there is a speed limit in the processing of queries on the power data. In this paper, we propose an architectural design of storage that has high processing speed of data streams in power data analysis. Our proposed storage system, which has high performance query processing of power data streams, has two features, i) multiple indices based storage schemes for fast retrieval and ii) high-speed query processing based on view materialization for iterative processing.