Rockset review: Real-time SQL for operational data


There are, or were, two kinds of SQL database—transactional (OLTP) and analytic (OLAP). After some technological advances, you could combine the OLTP and OLAP functionality and have a hybrid (HTAP). If you have an OLAP database that massively scales complex queries on historical data, then you have a data warehouse.


If you take out the transactions, add search, do schemaless ingestion, and do analytics on real-time data, then you have operational analytics. There’s currently only one database designed for operational analytics, Rockset.

It’s easy to confuse Rockset with a data warehouse, but they are used for somewhat different purposes. You want a data warehouse to analyze petabytes of historical data that you’ve ingested from your databases of record, and for the queries to run in minutes. You want Rockset to analyze gigabytes to terabytes of recent, real-time, and streaming data, and for the queries to run in milliseconds.