This topic describes the scenario for the data model test and the scenario for the performance metric test.

Data model test

The test data model simulates the data model of the electricity quantity data that is collected by a power grid. The data model contains the metrics that each electricity metering device reports at a specific point in time. Each electricity metering device has the following five tags:
  • sf: This tag indicates the province where the device is located.
  • ds: This tag indicates the city where the device is located. The value is the initials of the pinyin of the city name.
  • qu: This tag indicates the district where the device is located.
  • yxt: This tag indicates the type of the electricity quantity metering device. The value is TMR.
  • sbbh: This tag indicates the device ID. Each device has a unique device ID such as 10001.
Note The data reported by each device contains five metrics. These metrics are divided into two categories.
  • Active power: The pa and pb metrics are used to measure the active power.
  • Voltage: The va, vb, and vc metrics are used to measure the voltage.

Performance metric test

This section describes the performance metric values for writes and queries in the test.

  • Points per second (PPS) of writes

    Write PPS indicates the average number of data points that the electricity metering devices write to a database every second. To ensure the consistency of the measurement unit for the time series data that is written to the database, the number of single-value data points are collected in this test. For example, in the data model described in the preceding section, the time series data uses the single-value model. Each time a device reports data, the data contains the values of the five metrics. Therefore, five values are reported at a time. If the data is stored in a relational database table, the data reported at a time is stored in one row. Therefore, in this data model, five PPS in a time series database table is equal to one tuple per second (TPS) in a relational database table.

  • Queries per second (QPS)

    QPS refers to the average number of queries that are run per second.