Electricity retail cost calculation in conditions of dynamic pricing

The transition from a fixed tariff to dynamic pricing requires the modernization of the information infrastructure of an energy marketing company, when it is necessary to store an archive of electricity consumption for each device, averaged over certain time intervals. The purpose of the study is to substantiate modern tools for calculating the retail cost of electricity in the context of dynamic pricing and to formulate conditions and approaches to its improvement. As part of the study, a test bench was deployed on which the performance of the time series database (TSDB) was tested in comparison with SQL databases, as well as the operation of coroutines for computational operations. Experimental results have shown that MySQL provides 2 times higher speed when writing data compared to InfluxDB, however, with an increase in data volume, MySQL performance decreases due to the need to build indexes. InfluxDB, in turn, demonstrates stable write speed, but MySQL loses due to the peculiarities of the file architecture. Data reading analysis revealed the opposite trend: InfluxDB executes queries 8 times faster than MySQL, due to the optimized structure of time series storage and the ability to perform aggregate operations with less computational load. However, using InfluxDB is accompanied by a significant increase in processor load. However, using InfluxDB is accompanied by a significant increase in processor load. To optimize the calculation process, the segmentation of the database of readings by ranges of meter identifiers was implemented, as well as the use of parallel data processing using coroutines. These measures have significantly reduced the calculation time and improved the overall performance of the system. The developed architecture demonstrates the possibility of a scalable solution for automation of electricity metering, providing flexibility and a high degree of integration with existing accounting systems.