Sign in to save your place.

Meet Amazon Redshift

AWS's cloud data warehouse that powers analytics for thousands of companies. Let's understand what it is and why it matters.

☁️ What is Amazon Redshift?

Amazon Redshift is a fully managed, petabyte-scale cloud data warehouse by AWS. It's designed for running complex analytical queries on massive datasets — think billions of rows analyzed in seconds.

🧒 ELI5: The Giant Calculator in the Cloud

Imagine you have a super-powerful calculator that lives in Amazon's cloud. You can send it millions of math problems at once, and it solves them all in seconds. You don't need to buy the calculator — you just rent it by the hour. That's Redshift!

Unlike your laptop's database that struggles with millions of rows, Redshift can handle trillions of rows because it splits the work across many computers working together.

🏗️ How is Redshift Different?

FeaturePostgreSQLRedshiftSnowflake
TypeOLTP (transactions)OLAP (analytics)OLAP (analytics)
StorageRow-basedColumnarColumnar
ScaleSingle serverMulti-node clusterElastic cloud
PricingSelf-hostedPer-node/hour or ServerlessPer-credit
Best ForApp backendsAWS-centric analyticsMulti-cloud analytics
SQLStandard PostgreSQLPostgreSQL-compatibleANSI SQL

💡 Key Insight

Redshift is based on PostgreSQL 8.0, so if you know PostgreSQL, you already know ~80% of Redshift SQL! But Redshift adds columnar storage, MPP (Massively Parallel Processing), and deep AWS integration.

⚡ Why is Redshift Fast?

MPP
Massively Parallel Processing
Columnar
Column-oriented storage
Compressed
Automatic compression
Cached
Result caching

🧒 ELI5: The Team of Workers

Imagine you need to count all the red balls in a warehouse with 1 million balls. PostgreSQL sends ONE person to count them all — takes hours.

Redshift sends a TEAM of 100 workers, each counting their section simultaneously. Done in minutes! That's MPP — splitting work across multiple "nodes" (computers).

🎯 When Should You Use Redshift?

✅ Perfect For

  • You're already on AWS (S3, Glue, Lambda, etc.)
  • Business intelligence dashboards (Tableau, Looker, Metabase)
  • ETL/ELT pipelines with large datasets
  • Data science teams running complex queries
  • Predictable workloads with consistent cluster sizing

❌ Consider Alternatives When

  • Multi-cloud strategy → Snowflake or BigQuery
  • Real-time sub-second queries → ClickHouse
  • Transactional workloads → PostgreSQL/MySQL
  • Highly variable workloads → Snowflake (auto-scaling)

🏢 Who Uses Redshift?

Netflix
Streaming analytics
Airbnb
Booking analytics
Nasdaq
Financial data
Yelp
Review analytics

🧠 Quick Quiz

Redshift uses which type of storage?

October 2026: what to open first

A new warehouse is Redshift Serverless or an RA3 cluster. DC2 is a previous generation. The setup lesson is where that choice is made. Zero-ETL, data sharing, AUTO table design, automatic WLM, Spectrum, and Iceberg each have limits. The later lessons name the limit instead of treating the feature as always on.

Try this. Write one sentence: Serverless or RA3, and why you are not choosing DC2.

Knowledge check

What should a new Redshift warehouse use?

Course Hub Cluster Setup