TideDB - A Distributed, Scalable Time Series Database

Some of the largest datasets have strong time components, like machine monitoring, real-time alert and IoT devices, etc. Despite of so many applications of time series data, most storage options are either highly proprietary or worse, relational. Unlike other alternatives, TideDB does not have a data with multiple metrics broken down into multiple data with one metric that increases the pressure on system throughput dramatically, rather its data modeling based on the computed column and tag words index can provide high write throughput, low read latency, and petabytes storage. TideDB has been deployed in production settings on large clusters to manage multiple terabytes of storage at Taide Company. The paper describes the TideDB how to store and organize our time series data from about one hundred thousand devices and millions service modules.