Search documentation
karat

+

K

User Documentation ↗

Dataset basics

A Foundry stream is a combination of a streaming dataset (for "cold storage") and a buffer of data (for "hot storage"). Streams are particularly useful for workflows that require second-level latency and constant compute, such as real-time operational data analysis and processing.

Each branch of the streaming dataset can have one active stream.

For more information on streaming datasets, review Streams documentation.