Cloud Arena
Back to Data & Analytics

Amazon Redshift

aws

Fast, petabyte-scale cloud data warehouse using Massively Parallel Processing (MPP), RA3 managed storage, and Redshift Spectrum.

DATA & ANALYTICS
•Amazon Web Services

Fast, petabyte-scale cloud data warehouse using Massively Parallel Processing (MPP), RA3 managed storage, and Redshift Spectrum.

Click any section above to jump directly to it
GCP ➔ AWS Bridge
Fast-Track Mental Model Translation

Coming from Google Cloud?

Compare with Google Cloud BigQuery

Both are enterprise-grade cloud data warehouses, but BigQuery is 100% serverless on-demand, while Redshift offers dedicated provisioned MPP clusters and Redshift Serverless.

What's Conceptually Identical
  • Columnar storage engines optimized for large analytical aggregations and scans.
  • Decoupled compute and storage architectures (RA3 RMS vs. BigQuery Capacitor).
  • Direct SQL querying of external data lakes (Redshift Spectrum vs. BigQuery BigLake).
Key Architectural Shifts in AWS
  • Data Sharding: BigQuery manages slots automatically; Redshift requires defining `DISTKEY` and `DISTSTYLE` for optimal join efficiency.
  • Indexing: BigQuery uses table clustering; Redshift uses `COMPOUND` or `INTERLEAVED SORTKEY` zone maps.
  • Pricing: BigQuery defaults to $5/TB on-demand; Redshift offers fixed node-hour clusters and Redshift Serverless RPUs.

What is Amazon Redshift?

Amazon Redshift is a fully managed, petabyte-scale Massively Parallel Processing (MPP) relational data warehouse. A Leader Node compiles and optimizes SQL queries into high-performance C++ execution plans, while Compute Nodes execute tasks across dedicated hardware Slices in parallel. Modern RA3 instances decouple compute from storage using Redshift Managed Storage (RMS) backed by S3. Redshift Serverless automatically scales Redshift Processing Units (RPUs) dynamically, scaling to zero when idle.

Decision Guide: When to Choose This Service

When to Use Amazon Redshift

Use Amazon Redshift for enterprise data warehousing, high-concurrency SQL business intelligence dashboards, petabyte-scale joins, cross-cluster zero-copy data sharing, and when you need dedicated sub-second query performance across billions of rows.

When NOT to Use (Recommended Alternatives)

Do NOT use Amazon Redshift for operational OLTP transactional processing with frequent single-row inserts/updates (use Amazon Aurora instead) or simple serverless ad-hoc log querying (use Amazon Athena).

Core Architectural Pillars

MPP Architecture

Leader Node compiles C++ query plans; Compute Node Slices execute in parallel.

Redshift Managed Storage (RMS)

Automatically tiers hot working data to NVMe SSDs and cold data to Amazon S3.

Distribution Styles

`DISTSTYLE KEY` for local zero-network-shuffle joins; `DISTSTYLE ALL` for small lookup tables.

Sortkeys & 1MB Zone Maps

Columnar block min/max metadata eliminates disk I/O on date and numeric range filters.

Redshift Serverless

Automatic scaling of Redshift Processing Units (RPUs) with scale-to-zero capability.

Data Sharing

Zero-copy, live transactional table sharing across clusters, accounts, and regions with compute isolation.

Aurora Zero-ETL

Sub-minute transactional replication from Aurora to Redshift without custom Glue/Spark pipelines.