IOT & TELEMETRY

IoT & industrial telemetry

Millions of sensors, out-of-order arrivals, unbounded growth — handled by primitives built for exactly that shape of data.

THE STATUS QUO

Device data arrives late, duplicated, and forever. General-purpose databases buckle at the ingest rate; time-series silos cut you off from your reference data.

HOW EzraDB DOES IT

Device retries and at-least-once buses are deduplicated on DEDUP KEY (device_id, ts) — the table stays clean without app-side bookkeeping.

Align sensor A with the nearest reading from sensor B inside a tolerance window — the canonical sensor-fusion query, native in SQL.

Gorilla for floats, Delta for timestamps, RLE for states — sensor streams compress hard and scan fast.

Age raw readings out automatically while rollups stay hot; push cold history to object storage.

Reconstruct what the fleet looked like at the moment of an incident with AS OF SYSTEM TIME (row engine).

*In development — not yet production-ready.

IN PRACTICE

The sensor data stays queryable next to the reference data that gives it meaning.

sensor fusion, one clause
SELECT q.ticker, q.ts, t.price
FROM quotes q
ASOF JOIN fills t
  ON t.ticker = q.ticker AND t.ts >= q.ts   -- most recent fill at/after each quote
ORDER BY q.ts;