LoginSign Up
← Back to all sparks
A

Apache IoTDB

DEVOPS
Velocity0.0

Time series database for IoT and industrial data

iotdb.apache.org ↗

The table model is becoming a real SQL engine, and a C driver opens the industrial edge.

time-seriestable-modelsql-engineiotapachedual-branch
◆Current state
IoTDB runs two lines in parallel: 1.3.x carrying the original tree model and 2.0.x where nearly all new work lands. The 2.0 releases have been steadily building out the table model — set operations and common table expressions, window and pattern-recognition functions, JOIN variants including ASOF, approximate aggregates, user-defined table functions — turning what began as a time-series schema into something closer to a full SQL surface. Alongside that, an AINode component gained built-in forecasting models and inference for both models, and 2.0.10 added C-language driver SDK interfaces with parameter binding and multi-node failover.
◆Where it's heading
Two audiences are being served at once. The table model work courts analysts and existing SQL tooling, with Spark integration and Python DataFrame returns as the connective tissue; the C driver and failover handling court the embedded and industrial systems that generate the data in the first place. The 1.3 branch now receives only what can be backported — the March security hardening shipped to both lines with identical notes — which reads as a maintenance line with a finite life. Security posture also tightened noticeably in 2.0.7, which removed risky RPC interfaces and JEXL functions and changed default bind addresses to loopback.
◆Prediction
Expect continued SQL surface expansion in the table model and more client language coverage now that the C driver exists. The 1.3 branch's end is the open question these entries do not address.

◆Recent moves

  1. 1mo ago

    Set operations, CTEs, and a C driver SDK land in 2.0.10

    ⚡ SPARK

    2.0.10 brings set operations and common table expressions to the table model and ships C-language driver SDK interfaces with parameter binding and automatic multi-node failover. It advances both halves of IoTDB's strategy in one release — SQL completeness for analysts, native client access for the devices at the edge.

    View source ↗
  2. 3mo ago

    Query latency system tables and batched Python DataFrames

    Adds observability system tables for query latency and DataNode connection status, lets the Python client return results as batched DataFrames, and gives synchronization pipes finer path selection including device exclusions. Operational and integration polish on the 2.0 line rather than new query capability.

    View source ↗
  3. 5mo ago

    Risky RPC interfaces and JEXL removed; defaults bound to loopback

    A security hardening release that removes risky RPC interfaces and JEXL functions, changes the default client RPC address to loopback, and binds internal services to the configured internal address rather than a default. Reducing exposed surface and tightening default binds is the kind of change that breaks careless deployments and protects careful ones.

    View source ↗
  4. 5mo ago

    Security hardening backported to the 1.3 branch

    The same security hardening backported to the 1.3 branch with an identical change list, shipped a week ahead of the 2.0.7 release. It contains nothing beyond its 2.0 counterpart and confirms 1.3 is now maintained for security parity only.

    View source ↗
  5. 6mo ago

    Table model gains write-back and pattern-matching aggregates

    Extends the table model with query write-back, aggregate support in MATCH RECOGNIZE pattern matching, bit manipulation functions and a pushdown-enabled EXTRACT, while fixing three CVEs. Part of the sustained run of table model query work that 2.0.10 continues.

    View source ↗
  6. 7mo ago

    FastLastQuery interface and compaction efficiency gains

    A performance-oriented release on the 1.3 line: faster multi-series Last queries with a dedicated FastLastQuery SDK interface, streamed schema fetching for large datasets, and improved compaction and memory management. One of the last 1.3 releases with substantive new work rather than backports.

    View source ↗