OLAP vs OLTP
OLAP and OLTP are two database workload types. OLAP, Online Analytical Processing, is built for fast analytics over large datasets. OLTP, Online Transaction Processing, is built for fast reads and writes of individual records.
When to use each
OLTP is what runs your application. It handles many small, fast operations: create an order, update a balance, read one user's profile. It cares about consistency and quick single-row access. Postgres and MySQL are classic OLTP databases.
OLAP is what runs your analytics. It handles large scans and aggregations: revenue by region, sensor averages over a year, error rates by service. It cares about reading huge volumes fast. ClickHouse, Snowflake, and Arc are OLAP systems.
You almost always need both. The mistake is forcing one to do the other's job, like running heavy analytics on your production Postgres and watching it slow your app to a crawl. The clean pattern is OLTP for the application, OLAP for the analytics, with data flowing from one to the other.
How Arc handles OLAP vs OLTP
Arc is the OLAP half of that picture for time-stamped data. It takes the telemetry, events, and metrics your application generates and makes them queryable at scale, without putting analytical load on your transactional database.
Related terms
Arc is a high-performance columnar database. Open Parquet on storage you own, single Go binary, production-ready in 30 seconds.