In some ways we're similar to Redshift in others we're very different. At the core we both have Postgres as a foundation, though Redshift's roots in Postgres are from some time ago. Citus works with the latest Postgres release and is an extension which doesn't fork the underlying database. Both us and Redshift work with large scale data, from there we diverge a bit.

Redshift very much focuses on data warehousing, so commonly large batch loads and then complex queries and analysis which run over a longer period of time minutes and up commonly.

Citus focuses more on real-time analytics. We support real-time ingest so you can insert directly into it, or of course bulk load as well. And then on a query side because we have high parallelism customers typically are doing more with real-time analytics, so queries that operate in the second or lower range. An example is CloudFlare, which is a customer, uses us to power the dashboard you see when you login as a user.

A couple thoughts:

- With Redshift, you can just click a couple buttons in the AWS console to create a new DB and try it out. When you're done experimenting, you can shut it off.

- Citus is open source, so you can experiment on your own hardware.

- If you already have data as CSVs in an S3 bucket, it's a couple SQL commands to load it into Redshift. I don't see as nice an import option for Citus.

- Citus is designed to support mixing operational and analytic workloads. Redshift is just for analytics.

- Citus Cloud plans start at $1,800/month. Redshift clusters can be made a lot cheaper by snapshotting and spinning them down when not in use.