InfluxDB vs TDengine

Quick verdict

This page is for architects choosing a dedicated store for plant, building and energy telemetry: chillers, AHUs, meters and production equipment read every few seconds, feeding dashboards, alarms and maintenance systems. For background on how these stores fit between brokers and analytics platforms, see the time-series and IoT data hub.

InfluxDB 3 is InfluxData's current generation, written in Rust on Apache Arrow, DataFusion and Parquet. InfluxDB 3 Core is the open source release (MIT or Apache 2.0) aimed at recent data; InfluxDB 3 Enterprise adds historical queries, high availability and read replicas under a commercial licence. Older 1.x and 2.x releases (with Flux) remain documented. The latest InfluxDB release on GitHub when checked was v3.12.2. See the InfluxDB guide.

TDengine TSDB from Taos Data is written mainly in C and C++ and positioned for IoT, industrial internet, power and connected vehicles. TDengine TSDB-OSS is under the GNU Affero General Public License v3 and, according to the vendor, includes clustering; TDengine TSDB-Enterprise and TDengine Cloud add GUI connectors for MQTT and OPC, TLS on the REST and WebSocket endpoint, audit logs and other features. The latest release on GitHub when checked was 3.4.3.9 (3 November 2026). See the TDengine guide.

Side by side

Key differences

Data model: tags and fields versus one table per device

InfluxDB 3 groups similar points in a table, for example chiller, with tags such as site and chiller_id identifying the series and fields holding the values. Tags are part of the primary key and set the sort order of the Parquet files; the first write fixes the tag column order, and tag columns cannot be changed afterwards. The InfluxDB 3 schema guide says tag value cardinality no longer affects overall performance as it did in 1.x and 2.x, but a table has a maximum column count, so wide "one field per point" tables need care.

TDengine asks for a separate table for every data collection point, created from a supertable that acts as a template for one device type. Static attributes (site, plant room, model) are tags stored once per subtable and can be changed later; queries on the supertable filter subtables by tag and merge them. The vendor's reasoning is that each device then has one writer, so writes are lock-free and append-only, and its data is stored contiguously. Plan names, tag sets and the number of vnodes early: each table is assigned to a vnode by hashing its name at creation.

# Illustrative, not run here: line protocol format from the InfluxDB 3 Core documentation.
# TDengine also accepts this format in schemaless mode.
chiller,site=campus-a,chiller_id=ch-1 kw=412.0,chw_supply_c=6.8 1794042000000000000

-- Illustrative DDL, not run here: TDengine supertable and subtable, adapted from the
-- TDengine documentation; check it against the documentation for your version.
CREATE STABLE chiller (ts TIMESTAMP, kw FLOAT, chw_supply_c FLOAT)
    TAGS (site VARCHAR(32), plant_room VARCHAR(32));
CREATE TABLE ch_1 USING chiller (site, plant_room) TAGS ('campus-a', 'pr-1');

Querying: DataFusion SQL and InfluxQL versus windowed SQL

InfluxDB 3 uses Apache DataFusion SQL with DATE_BIN for bucketing and date_bin_gapfill with interpolate or locf for gaps, and keeps InfluxQL for 1.x-style dashboards. Flux is in maintenance mode and not supported in InfluxDB 3, so 2.x Flux code needs rewriting.

TDengine adds window clauses that suit equipment data: INTERVAL (fixed or sliding), SESSION (gaps), STATE_WINDOW (runs of the same state, such as a compressor on or off), EVENT_WINDOW (start and end conditions) and COUNT_WINDOW, with FILL and PARTITION BY tbname to aggregate each device separately. A per-chiller 15-minute average in each:

-- Illustrative, not run here: InfluxDB 3 SQL pattern from the InfluxDB 3 Core documentation.
SELECT DATE_BIN(INTERVAL '15 minutes', time, '2026-01-01T00:00:00Z'::TIMESTAMP) AS time,
       chiller_id, avg(kw) AS avg_kw
FROM chiller
WHERE time >= now() - INTERVAL '1 day'
GROUP BY 1, chiller_id
ORDER BY 1;

-- Illustrative, not run here: TDengine pattern adapted from its documented
-- PARTITION BY tbname INTERVAL example.
SELECT _wstart, tbname, AVG(kw) AS avg_kw
FROM chiller
WHERE site = 'campus-a' AND ts >= NOW - 1d
PARTITION BY tbname
INTERVAL(15m) FILL(NULL);

Ingestion, processing and getting data out

InfluxDB takes line protocol over HTTP and is compatible with the 1.x and 2.x write APIs, so Telegraf and existing clients work; Telegraf's MQTT consumer input subscribes to a broker. Transforms, downsampling and alerts run in the embedded Python processing engine, whose plugin library includes a downsampler and an InfluxDB to Iceberg plugin. A point is identified by table, tag set and timestamp; when duplicate writes set the same field, the documentation says InfluxDB does not guarantee which value is kept.

TDengine takes SQL INSERT (creating subtables on first write) or schemaless writes in InfluxDB line protocol and OpenTSDB formats, so the same Telegraf output can feed it. Its SQL streams transform data as it arrives, and data subscription exposes topics through a Kafka-like consumer API (and since 3.3.7.0 to MQTT clients). The no-code MQTT, OPC and historian connectors configured in taosExplorer are Enterprise features; with OSS you bridge the broker yourself. In either case, a broker such as Kafka or MQTT moves messages, the database stores and queries them, and CMMS records stay in a relational database.

Architecture, scaling and retention

InfluxDB 3 Core is a single node aimed at real-time monitoring and recent data, persisting Parquet to local disk or object storage. High availability, read replicas, multi-node clustering and historical query capability are InfluxDB 3 Enterprise features, and InfluxDB Clustered and the cloud services are the larger options. Retention is per database, does not expire by default, and in Core can be set only when the database is created, so raw and downsampled data usually go into separate databases.

TDengine OSS runs as a cluster of dnodes (taosd processes) holding vnodes, which shard a database's tables, with up to three mnodes for metadata replicated with Raft. Databases set REPLICA to 1 or 3 in OSS (2 is Enterprise only), KEEP for retention and DURATION for the span of each data file; KEEP must be at least three times DURATION, can be adjusted later, while DURATION cannot. Multi-tier storage and RSMA downsampled storage are Enterprise features. See retention and downsampling.

Editions, licences and security

InfluxDB. The influxdb repository states that its open source software is "licensed under the permissive MIT or Apache 2 licenses at the user's choosing". InfluxDB 3 Enterprise runs under a licence file from InfluxData (Trial, At-Home and Commercial types, limited by CPU count). Core uses admin tokens that grant access to every action on the server; per-database resource tokens are an Enterprise feature, so keep Core on a private network.

TDengine. The repository README states that TDengine "is licensed under the GNU Affero General Public License 3.0". Section 13 of the AGPL requires that, if you modify the program, your modified version must offer users interacting with it over a network the opportunity to receive its corresponding source. TLS for the public WebSocket and REST endpoint, IP allow-lists, audit logs and encryption at rest are documented as Enterprise features, so with OSS keep the cluster on a private network or terminate TLS at a proxy. Not legal advice: this summarises the licence texts; whether the AGPL affects your use is a question for the licence and your advisers.

Pricing and licensing

InfluxDB. InfluxData's pricing page states that InfluxDB 3 Core "is open source and free to download and run with no license required". InfluxDB 3 Enterprise is priced by CPU configuration through sales, with a 30-day free trial. Cloud Serverless is usage-based (data in, queries, storage and data out) with a free tier limited to 30-day retention; the page lists unit prices without stating a region, so we do not quote them. Cloud Dedicated is custom-priced.

TDengine. TDengine TSDB-OSS (AGPL v3) carries no licence fee. The vendor's pricing page lists the commercial TDengine All-in-One product, which includes TDengine TSDB and TDengine IDMP, priced by tags ("a unique time-series signal, typically Asset/Device + Measurement Name"), with "All prices are in USD". Self-hosted is "free to use for up to 5,000 tags" (optional support USD 3,375 a year); for example 10,000 tags is listed at USD 15,500 a year or USD 38,750 perpetual. The fully managed cloud is listed at USD 970 a month for 5,000 tags and USD 1,740 a month (USD 20,880 a year) for 10,000 tags, on AWS, Azure or Google Cloud. High availability adds 30% to self-hosted plans and 50% to cloud plans; very large or unlimited-tag deployments are quoted. No tax or region note was given.

Pricing checked on the vendors' official pages on 7 November 2026. Prices change; confirm before buying.

Where each one leads

Limitations

When to choose each

When neither is right

• Readings must be joined to assets and work orders in SQL: see TimescaleDB vs InfluxDB and QuestDB vs TimescaleDB.
• The data is infrastructure metrics for PromQL dashboards and alerts: see InfluxDB vs Prometheus and VictoriaMetrics vs Prometheus.
• You want an Apache-licensed engine with a device model: see the Apache IoTDB guide.
• You are still shortlisting: see the time-series database selection guide.

Final recommendation

Frequently asked questions

Sources

• InfluxDB 3 Core documentation
• InfluxDB 3 Core: get started and data model
• InfluxDB GitHub repository and releases
• InfluxData pricing
• TDengine documentation: Introduction
• TDengine documentation: Basic concepts
• TDengine documentation: Architecture
• TDengine documentation: Time-series extensions
• taosdata/TDengine releases on GitHub
• TDengine All-in-One pricing

Checked November 2026.

How we research comparisons: our editorial method.