Amazon Redshift

Overview

Amazon Redshift is a cloud data warehouse from AWS, built for analytical (OLAP) SQL workloads over large datasets. It uses a massively parallel processing (MPP) architecture: data is distributed across nodes, and queries run in parallel across them. Redshift's MPP engine has roots in ParAccel technology, and its SQL front-end is derived from PostgreSQL, so its dialect and client connectivity are largely PostgreSQL-compatible even though the storage and execution engine are proprietary and columnar.

Deployment options

Redshift can run in two ways:

• Provisioned clusters — you choose node types and cluster size, giving direct control over compute.
• Redshift Serverless — capacity is allocated automatically, so you run queries without managing a cluster.

Related features include Redshift Spectrum for querying data in Amazon S3 without loading it first.

Redshift and DuckDB

Redshift is a distributed, managed warehouse designed to scale across many nodes inside AWS. DuckDB is an open-source, single-node, in-process OLAP engine, with MotherDuck as its serverless cloud version. For many analytical workloads that fit on one machine, DuckDB delivers strong columnar performance without cluster management; Redshift is aimed at centralized, elastically scaled cloud warehousing. Because Redshift exports to S3 and speaks SQL, DuckDB is often used alongside it for local development or ad-hoc file querying:

Copy code
-- DuckDB querying Redshift UNLOAD output in S3
SELECT customer_id, SUM(amount) AS total
FROM read_parquet('s3://warehouse/unload/orders/*.parquet')
GROUP BY ALL
ORDER BY total DESC
LIMIT 10;

A comparable Redshift query, using its PostgreSQL-style dialect:

Copy code
SELECT customer_id, SUM(amount) AS total
FROM orders
GROUP BY customer_id
ORDER BY total DESC
LIMIT 10;