TDengine
activeby TAOS Data · since 2019 · also known as TDengine TSDB, TDengine OSS, TAOS
Industrial-IoT time-series database with a table-per-device model and built-in streaming and caching.
Overview
TDengine, from TAOS Data, is a distributed time-series database designed around industrial IoT. Each device gets its own sub-table under a shared "super table" schema, which keeps writes sequential and per-device queries fast, while tags on the super table drive fleet-wide aggregation. It bundles features other stacks bolt on, including stream processing, Kafka-style data subscription, last-value caching and retention, so one cluster can stand in for a broker, a TSDB and a cache for sensor data. The open-source edition is AGPL-3.0 and includes clustering; TDengine Enterprise and TDengine Cloud add tiered storage, backup and connectors to industrial sources such as OPC and PI.
Strengths & trade-offs
Strengths
- +Super-table model fits fleets of millions of devices sharing a schema
- +Built-in stream processing, data subscription and last-value cache reduce extra infrastructure
- +Columnar per-device storage gives high compression on sensor data
- +Accepts InfluxDB line protocol and OpenTSDB writes in schemaless mode
- +Clustering is included in the open-source edition
Trade-offs
- −AGPL-3.0 licence is a hurdle for embedding in commercial products
- −SQL dialect has restrictions (joins mainly on timestamps, limited updates)
- −Community and docs are strongest in China; English ecosystem is smaller
- −Table-per-device design suits regular telemetry better than ad-hoc event data
Great for
- +Industrial IoT and SCADA historian replacement
- +Connected-vehicle telemetry
- +Smart meters and energy grids
- +Edge-to-cloud sensor data aggregation
- +Real-time equipment dashboards
Think twice for
- −General OLTP workloads
- −Log search
- −Irregular event data without a stable schema
Capabilities
Native or official first-party support · 10 of 25 tracked
Scale & consistency
- Comfortable data size
- Large · ≤ 1 PB
- Horizontal scaling
- Native
- Multi-region
- No
- Transactions
- None
- Consistency
- Strong
- Schema
- Strict
- Architecture
- Distributed, Single node, Client server
- Query languages
- SQL
Data is split into virtual nodes (vnodes) replicated with Raft and spread across dnodes.
Deployment
Ecosystem
Alternatives to TDengine
Apache time-series database for industrial IoT with tree and table models and edge-to-cloud sync.
tdengine vs apache-iotdb →Purpose-built time-series database, rebuilt in Rust on Arrow, DataFusion and Parquet for version 3.
tdengine vs influxdb →PostgreSQL extension adding hypertables, columnar compression and continuous aggregates for time series.
tdengine vs timescaledb →Columnar time-series SQL database built for market data and very high ingest rates.
tdengine vs questdb →Rust cloud-native time-series database with SQL on DataFusion and InfluxDB-style ingestion.
tdengine vs cnosdb →Toshiba's memory-first container database for IoT time series, with per-container ACID and SQL.
tdengine vs griddb →TDengine FAQ
What is TDengine?
TDengine is a time-series database by TAOS Data, first released in 2019. TDengine, from TAOS Data, is a distributed time-series database designed around industrial IoT. Each device gets its own sub-table under a shared "super table" schema, which keeps writes sequential and per-device queries fast, while tags on the super table drive fleet-wide aggregation. It bundles features other stacks bolt on, including stream processing, Kafka-style data subscription, last-value caching and retention, so one cluster can stand in for a broker, a TSDB and a cache for sensor data. The open-source edition is AGPL-3.0 and includes clustering; TDengine Enterprise and TDengine Cloud add tiered storage, backup and connectors to industrial sources such as OPC and PI.
Is TDengine open source?
Yes. TDengine is open source. License: GNU AGPL v3.0 (open-source edition).
Does TDengine support ACID transactions?
No. TDengine does not provide ACID transactions; its consistency model is strong.
Does TDengine scale horizontally?
TDengine scales horizontally out of the box, with built-in sharding or partitioning. It is comfortable at large scale (≤ 1 PB).
How can I run TDengine?
You can self-host it, use the vendor's managed cloud and run it at the edge. Managed offerings include TDengine Cloud.
What is TDengine best for?
TDengine is a strong fit for industrial IoT and SCADA historian replacement; connected-vehicle telemetry; and smart meters and energy grids. Think twice for general OLTP workloads.
What are the best alternatives to TDengine?
Common alternatives to TDengine include Apache IoTDB, InfluxDB, TimescaleDB, QuestDB and CnosDB. The right choice depends on your workload, scale and hosting; WhichDB's smart match ranks them against your requirements.
How popular is TDengine?
TDengine ranks #192 of 844 databases on the WhichDB Score (17.4/100), and #10 among time-series databases. The score combines public adoption signals such as package downloads, GitHub activity, job posts and developer discussion.