AWS Certified Cloud Practitioner / Databases & Analytics

Redshift Overview

📊 Amazon Redshift

Storing data isn't enough — organizations need to analyze it for trends and decisions.

Redshift is AWS's fully managed data warehouse, built for analytics and BI on large datasets — unlike RDS, which handles everyday app transactions

🎯 Exam Tip

  • ✅ RDS = OLTP (application databases)

  • ✅ Redshift = OLAP (analytics/data warehousing)

  • ✅ Redshift uses columnar storage

  • ✅ Common BI pairing: Tableau, QuickSight

  • ✅ Redshift Serverless auto-manages infra and scales on demand

🧠 Core Idea: OLTP vs OLAP

An online store runs constant small transactions all day — signups, logins, orders, cart updates, payments. That's OLTP (Online Transaction Processing) — exactly what RDS is built for.

Now management asks: Which products sold best this month? Which country made the most revenue? What are the holiday trends?

Answering that means scanning huge volumes of historical data — OLAP (Online Analytical Processing). That's what Redshift is built for.

🏢 What is a Data Warehouse?

A centralized repository for large volumes of historical data from multiple sources — built to store massive datasets, run complex analytical queries, generate reports, and power dashboards/BI tools.

🎯 Exam tip: "Analytics," "reporting," "dashboards," "data warehousing" → Redshift

Unlike transactional databases, Redshift doesn't ingest every transaction live — data loads in batches (hourly, every few hours, or nightly), then gets analyzed.

Traditional databases (like RDS/PostgreSQL) store data row by row. Redshift stores it column by column instead.

Why it matters: need the average of the Sales column? Redshift reads just that column instead of scanning every field in every row — much faster for analytical queries.

🎯 Exam tip: "Columnar storage" → Redshift

Instead of one computer chewing through a huge query, Redshift splits the work across many compute nodes at once — like 100 people each counting 10,000 records instead of one person counting a million. Much faster on large datasets.

Fast analytical queries, scales to petabytes, and — since Redshift is PostgreSQL-based — you query it with standard SQL.

Redshift plugs into BI tools like Amazon QuickSight and Tableau, turning raw data into charts and dashboards for tracking trends and business performance.

Standard Redshift is provisioned — you choose cluster size/compute in advance and pay for what you provisioned.

🚀 Amazon Redshift Serverless

AWS automatically manages infrastructure, capacity planning, scaling, and provisioning — you just run queries.

⚙️ How It Works

No servers or clusters to manage — AWS handles it all.

🎯 When to Use It

Great for reporting, dashboards, real-time analytics, ad hoc SQL, and workloads where demand is unpredictable — AWS scales automatically as needed.

📌 Redshift vs Redshift Serverless

Redshift

Redshift Serverless

You provision and manage the warehouse

AWS provisions and manages it automatically

Pay for provisioned compute

Pay only for what you use

Best for predictable workloads

Best for variable/unpredictable workloads

📝 Summary

  • 📊 Redshift = AWS's fully managed data warehouse

  • 🧠 Built for OLAP, not transactional workloads

  • 📈 Use for analytics, reporting, BI, and data warehousing

  • 🗂️ Columnar storage makes analytical queries fast

  • ⚡ MPP processes massive datasets efficiently in parallel

  • 📊 Integrates with QuickSight and Tableau

  • 🚀 Redshift Serverless = auto-managed infra, auto-scaling, pay-per-use