Cloud Parts

TimechoDB

The commercial edition of Apache IoTDB from the company its creators founded: the same engine plus dual-active pairs, one-way data-diode sync, a Workbench console and vendor support, licensed per machine with activation codes.

Type
Commercial product
Made by
Timecho Limited
License
Proprietary, activated per machine with codes from Timecho (built on Apache 2.0 IoTDB)
Facts checked
9 October 2026

What it is

TimechoDB is Apache IoTDB as packaged and supported by Timecho, a company the IoTDB core team from Tsinghua University founded in October 2021. Under the hood it is the same system: ConfigNodes and DataNodes, data in Apache TsFile files, tree and table models, and AINode for forecasting from SQL with built-in models such as Timer-XL, Sundial and Chronos-2. Version numbers follow Apache’s with a fourth digit for Timecho’s builds, and those builds have shipped first: TimechoDB 2.0.10.1 came out on 25 June 2026 and Apache IoTDB 2.0.10 on 8 July; 2.0.11.1 on 11 August and Apache 2.0.11 on 11 September. The 1.x builds were called IoTDB Enterprise (iotdb-enterprise-1.3.x) and still get fixes, the latest being 1.3.7.3 in June 2026. Timecho says it releases every two to three months and offers day-level fixes for urgent customer issues.

What the docs confirm it adds: dual-active deployment, where two independent servers or clusters accept writes and mirror each other through pipes until both hold the same data, giving high availability with fewer than three machines (the docs call it an enterprise-only feature); an iotdb-air-gap-sink that syncs data across one-way data diodes such as the Nanrui Syskeeper 2000, which Apache’s sync docs don’t list; and tools: the Workbench console, a Grafana monitoring panel and a cluster deployment tool. Timecho’s own comparison table also marks tiered storage, data subscription, views, whitelists and audit logs as TimechoDB-only, but Apache’s docs describe data subscription since 1.3.3 and a tier_ttl_in_ms setting that moves data to the next storage tier, so read that table as partly out of date.

There is no public download. You apply for a trial or contact Timecho for the timechodb-<version>-bin.zip package, start it, run show system info to get a machine code, and send that code and the version to Timecho, which returns an activation code to enter with activate '...' or a licence file to drop into the activation folder. Since 2.0.11.2 codes are tied to a release version. Docker images come the same way. Timecho’s pages publish no prices. Its other products are Workbench and TimechoAI, a hosted forecasting service built on the Timer models from Tsinghua. The company is based in Beijing and has a European arm, Timecho Europe GmbH, in Stuttgart. Timecho says it serves more than 1,000 industrial companies in over 50 countries; the customers it names include Changan Automobile, CRRC Sifang, China Baowu Steel and CNNC’s nuclear plants.

Why teams pick it

  • High availability with two machines: a dual-active pair keeps serving when one side stops and catches the other up when it returns, while Apache IoTDB’s docs offer only a standalone server or a cluster of at least three machines
  • Built for segmented industrial networks: an air-gap sink pushes data across one-way data diodes, and pipes aggregate data from sites to a central system
  • Support from the people who write the code: Timecho engineers include Apache IoTDB and TsFile PMC members, and the Apache project lists Timecho Europe GmbH as a commercial support provider
  • Same data format and models as Apache IoTDB: TsFiles load into and out of TimechoDB, it has the same tree and table models, and each build sits on a matching Apache version line

Watch out for

  • Closed distribution: packages, Docker images, the monitoring panel and Workbench all come from Timecho, every node needs an activation code tied to its machine and, since 2.0.11.2, to the release version, and there is no public price list
  • Since 2.0.8.2 the package leaves out the MQTT and REST service JARs; you ask Timecho for them
  • The docs still recommend installing on physical or virtual machines rather than in Docker
  • Timecho’s feature comparison with Apache IoTDB is partly stale (Apache documents data subscription and storage tiers too), so check each claimed extra before paying for it
  • Its references are mostly in China: the named customers are largely Chinese industrial firms, and the docs and product pages stress compatibility certifications for Chinese CPUs and operating systems

How it fits with other technologies

What TimechoDB adds, checked against the Apache docs#

Timecho’s docs list features that TimechoDB has and Apache IoTDB lacks. Some of those claims hold; others are out of date.

Timecho saysWhat Apache IoTDB’s docs show
Dual-active deploymentTimechoDB only: two servers or clusters that both accept writes and mirror each otherNot offered: Apache’s deployment docs list only standalone (one machine) and cluster (three or more) modes
Data syncReal-time and file sync with rich built-in plugins; Apache has file sync onlyApache has real-time pipes with built-in Thrift and SSL sinks; the iotdb-air-gap-sink for data diodes appears only in Timecho’s docs
Data subscriptionTimechoDB onlyDocumented in Apache IoTDB since 1.3.3
Tiered storageTimechoDB only: hot, warm and cold tiers on SSD, disk or cloud storageApache’s configuration has tier_ttl_in_ms, which moves data to the next tier once it passes an age
ToolsWorkbench console, cluster management tool and monitoring panelNot shipped by Apache; the Workbench and monitoring panel are obtained from Timecho
Getting itTrial application or contact with Timecho; activation code per machineFree download under Apache 2.0, no activation
  • Timecho also claims more than ten times better performance than the open-source version with its support. That is a vendor claim with no published method.

See it in a real stack

No architecture uses TimechoDB yet. Drop it into the playground and wire it up yourself.Open playground →
Where it came from2021–2026

IoTDB’s creators trace it to 2011, when a Tsinghua University team met the data problems of Sany Group’s equipment and saw where traditional databases fell short for industrial IoT. They started IoTDB and its TsFile format in 2015, took it to the Apache Incubator in 2018, and it became a top-level Apache project in 2020.

In October 2021 the core team founded Timecho (天谋科技) through Tsinghua’s technology-transfer process, and in June 2022 it raised more than US$10 million in a first round led by Sequoia China, with Koalafund, Gobi China and Cloudwise. The founder, Xiangdong Huang, was also chair of the Apache IoTDB project, and Julian Feinauer, a PMC member, was its technical director for Europe. Timecho said it would put the money into open-source R&D, to meet enterprises’ specific requirements and add features such as end-edge-cloud collaboration.

The commercial builds first shipped as IoTDB Enterprise; the 2.0 line of January 2025 was named TimechoDB and brought the table model. A TimechoDB 2.0.2.1 system took first place on the TPCx-IoT version 2 performance list in 2025. At its August 2026 summit in Beijing, Timecho presented a stack of TimechoDB for multimodal time-series data, the TimechoAI model service and TimechoAgent for automation through a CLI and MCP.

Why it’s built this way

Sell an edition on top of the Apache project instead of forking it
Timecho’s builds follow Apache IoTDB’s version lines and its engineers remain PMC members of IoTDB and TsFile. The company sells extras, tools and support around the shared core, and its 2022 funding announcement named open-source R&D as the use of the money.
Dual-active pairs as well as Raft clusters
An Apache cluster needs at least three machines. Two independent machines that mirror each other give high availability with fewer than three, which the docs pitch as cheaper to deploy, and putting each on its own power and network ring keeps one failure from taking down both.
Sync across one-way data diodes
Energy customers move data between security zones through one-way network gateways, which Timecho’s product page lists as a requirement in power. A sink that writes through diodes such as the Nanrui Syskeeper 2000 lets that data reach a central system.
Models inside the database
AINode runs forecasting and anomaly detection on data where it is stored, so it doesn’t have to be copied to a separate machine-learning platform, which the docs say speeds processing, improves security and cuts cost.

Timeline

  1. 2021

    Timecho founded in October by the Apache IoTDB core team

  2. 2022

    Raises over US$10 million in June, led by Sequoia China

  3. 2023

    IoTDB Enterprise 1.1.0.1, the oldest build in Timecho’s release history, ships in April

  4. 2024

    IoTDB Enterprise 1.3.0.4 adds the AINode machine-learning framework in January

  5. 2025

    TimechoDB 2.0.1.2 brings the table model in January; a 2.0.2.1 system tops the TPCx-IoT V2 list

  6. 2026

    TimechoDB 2.0.11 moves to JDK 17; Timecho shows its DB × AI stack at its August summit

Sources23 · facts checked 9 October 2026

Improve this page on GitHub