Load Data
AetheriusDB gives you several ways to get data in, from a single INSERT to a
continuous multi-million-row stream. They all land in the same tables and are
queryable the moment they commit — pick the path that matches your workload.
Why it matters
Section titled “Why it matters”Ingestion is where most analytical stacks bolt on a second system — a streaming service, a CDC pipeline, a bulk loader, a connector fleet. AetheriusDB folds all of those into the database, so you deploy one thing, monitor one surface, and never stand up a separate tier just to move rows.
- One system, not a pipeline — streaming, bulk, CDC, and cloud connectors are built in; there’s no Kafka-in-the-middle or connector pods to operate.
- Queryable the instant it lands — a row is visible to readers the moment it commits; there is no “wait for the loader” gap.
- Writers never block readers — heavy analytical queries keep running at full speed while millions of rows stream in alongside them.
How fast it is
Section titled “How fast it is”- Streaming: acknowledged in microseconds. Rows are accepted into memory and acked immediately, then organized into columnar storage in the background — so a continuous stream ingests at line rate (whatever the socket can deliver) with no write-side stalls.
- Bulk: network-bound, not database-bound. The Apache Arrow / Flight SQL path
streams columnar data with no per-row serialization tax, so a
COPY FROMS3 is limited by your hardware, not the engine. - CDC: sub-second freshness. A committed row on a mirrored Postgres or MySQL lands in AetheriusDB within low-millisecond replication lag.
Choose a path
Section titled “Choose a path”| You want to… | Use | Page |
|---|---|---|
| Stream rows continuously with immediate acknowledgement | A continuously-ingested table | Continuous ingestion |
| Load large columnar datasets (Arrow / Parquet) at line rate | Arrow Flight SQL or COPY FROM | Bulk ingestion (Arrow) |
| Understand how the paths fit together | — | Native ingestion |
| Mirror a live Postgres or MySQL database | CREATE BRIDGE | CDC bridge |
| Pull from Kafka, S3, GCS, Azure, or HTTP | CREATE SOURCE | Source connectors |
| Monitor throughput, lag, and row counts | System catalog views | Producer & ingest stats |
How to read these pages
Section titled “How to read these pages”Every page carries an availability badge — Stable (on by default), Opt-in
(set a flag), or Beta (works, surface still evolving). The high-throughput
bulk path is opt-in behind BULK_PULSE_V2=1; see
Configuration & Availability for the authoritative
flag list.