PROVIDER Β· Chinese time-series database and IoT data platform
TDengine
TDengine is the Chinese time-series database built for IoT, where one table per device with supertables aggregates massive fleets in one SQL and high compression with tiered storage cuts IoT storage cost sharply.
4pts
Pros
Cons
Product Information
Detailed Notes
Overview
TDengine was founded in 2017 by TAOS Data as a flagship Chinese time-series database, using a model of one table per device plus supertables that separates tags from time-series data, with the core open sourced since 2021 alongside enterprise and cloud offerings.
For Backend & Database work the value is compression and querying tailored to IoT, since one device's readings sit physically adjacent for high columnar compression, and supertables manage massive device fleets by tags so one SQL aggregates across devices, with built-in downsampling and interpolation speeding up industrial dashboards.
Assess ecosystem and migration, because the SQL dialect differs from InfluxDB and TimescaleDB so queries change on migration, overseas community and English material are smaller, and advanced capabilities such as clustered HA and tiered storage are enterprise features with a limited open source edition.
Core Strengths
- One table per device plus supertables aggregates massive fleets in a single SQL
- High columnar compression with vendor claims of storage near a tenth of general databases
- Built-in downsampling, interpolation, and time-window functions speed up industrial dashboards
- Native edge-cloud collaboration and tiered storage fit industrial and energy deployments
Product Ecosystem
Supertables and tags
A supertable defines a device type with child tables per device and tags for dimensional filtering and grouped aggregation.
Ingestion and querying
SQL inserts plus multiple collection protocols, with downsampling, interpolation, and sliding windows in queries.
Deployment
Open source single node and cluster, enterprise HA clusters, and TDengine Cloud hosting.
Drawbacks
- The SQL dialect differs from mainstream engines, requiring query rewrites on migration
- The open source edition limits clustered HA and advanced features
- Smaller overseas community and English material for international projects
- Poor tag design hurts compression and query performance, needing planning
Use Cases
- Industrial equipment and energy monitoring (β β β β β ) - Supertables with tiered storage
- Massive IoT device data (β β β β β ) - One table per device
- Edge-cloud collaboration (β β β β ) - Native edge support
- International multi-cloud standards (β β ) - Less overseas material
Pricing
| Plan | Price | Scope |
|---|---|---|
| Open source | $0 | AGPL licensed with single node and basic clusters |
| Enterprise | Per node subscription | HA clusters, tiered storage, support |
| TDengine Cloud | Per instance and usage | Managed with elastic scaling |
The AGPL open source edition is free for most projects though distribution obligations apply, while enterprise subscriptions per node buy HA clustering, tiered storage, and support, with storage savings typically substantial against disk or cloud spend.
FAQ
-
TDengine or InfluxDB? TDengine for domestic IoT projects with massive device counts and low storage cost, InfluxDB for a mature overseas ecosystem and rich plugins.
-
What is a supertable? It unifies the schema of similar devices with tags distinguishing them, so SQL against the supertable aggregates across devices without naming each table.
-
Does AGPL matter? Internal use carries no extra obligations, while offering modified code as a service must comply with AGPL terms and closed distribution should buy a commercial licence.