A novel InfluxDB-based inverted index method

High performance time series database supporting petabytes of data per day; however, the database cannot search the log's text content. Therefore, this paper proposes an inverted index method for the InfluxDB time series data to improve the database's text data searching ability. The method comprises two index levels: the first level uses lock-free hopping tables to screen out potential time series data, and the second utilizes an inverted log content index to accelerate specific time points data within the series. Furthermore, the experiments demonstrate that the modified InfluxDB method possesses a text matching function that increases the log text's search speed by 13∼15.6 times that of the InfluxDB while ensuring the ability to read and write time series data remains unchanged.