MICROSERVICE OF FORECASTING MODELS CLOUD STORAGE AND DATA EXCHANGE AT THE DECISION SUPPORT SYSTEM FOR HYBRID POWER NETWORK MANAGEMENT

Abstract

The purpose of the study is to develop a microservice for storing forecasting models used in the decision support system for energy management of hybrid power grids and their metadata. This will help organize forecasting models when they are stored and increase the efficiency of managing forecasting models in cloud storage. Methodology. This work uses the methodology of system analysis, design of information systems, organization of data and file storage in cloud storage. The prototyping method was used for the development and testing of the microservice, namely the software implementation of the microservice prototype based on the RESTful API. Microservice deployment was done using containerization through Docker. The scientific novelty of the study consists in the development of a new architectural solution for the storage and management of prediction models placed in the S3 cloud storage. The architecture of the storage and data exchange subsystem was designed in the form of a microservice using the RESTful architectural template. Conclusions. The following results were achieved. The S3 object storage of forecasting models have been developed, which corresponds to the file storage architecture of cloud providers. A microservice has been developed that implements the API interface for processing requests to the storage of prediction models and provides management of the forecasting models stored in the storage. The developed microservice will be integrated with the decision support system for power grid management and can be used both locally and deployed on the platforms of leading cloud service providers. The microservice was tested using neural network forecasting models of various types of electricity consumption, which confirmed its functionality.