AWS · Software · YoctoIT tech page

Redshift

AWS's data warehouse: petabyte-scale analytics, integrated with the lake on S3 and now serverless too.

FOCUS · ANALYTICS ON THE CLOUDThe warehouse that queries the lake too: SQL on S3 without moving the data
Request the Free Assessment Learn more
YoctoIT material for clients and partners · AWS e i nomi dei servizi sono marchi di Amazon.com, Inc.
01 · What it is

Amazon Redshift, made clear.

Redshift brings the data warehouse to the AWS cloud: compressed columns, massively parallel execution and the ability to query data on S3 directly (Spectrum). The Serverless version removes cluster management too.

Petabyte
the warehouse scale, without appliances
Serverless
you pay for queries, not for the running cluster
Zero-ETL
from Aurora and RDS to the warehouse without pipelines to maintain
Icona ufficiale AWS — Amazon Redshift
OFFICIAL AWS ICON · Redshift
Console AWS
REAL AWS CONSOLE · REDSHIFT QUERY EDITOR V2 · SOURCE: AWS BLOG
02 · How to use it well

The things that make the difference.

The analytics platform

BI & data appsLooker, Power BI, QuickSight
Redshift
Spectrum → S3
Data sharing
warehouse · lake · sharing
Zero-ETL from RDS/Aurorathe transactional data that arrives by itself
S3 data lakethe common base of the data
From raw data to the dashboard

Columnar MPP architecture

Analytic queries over billions of rows: compression and parallelism do the work.

Redshift Spectrum

SQL that reaches into the lake: S3 queried without loading anything.

Materialized views & autotuning

Aggregations that maintain themselves: fast BI without endless tuning.

Data sharing

Data shared across environments and group companies without copies: a single truth.

03 · In depth

MPP, RA3 and the warehouse that stops

Redshift is a columnar MPP: tables are distributed (DISTKEY/EVEN/ALL) and sorted (SORTKEY) for join co-location; RA3 nodes separate compute and managed storage, Serverless bills in RPUs only when querying. Spectrum reads external S3, materialized views speed up dashboards, WLM/queues manage concurrency, and zero-ETL from Aurora brings transactional data in without pipelines.

  • DISTKEY/SORTKEY — the physics of data: co-located joins are worth more than any tuning
  • RA3 — compute and storage separated: you scale compute without redistributing data
  • Serverless — RPUs on consumption: the warehouse for peaks and non-24/7
  • Spectrum — direct queries on S3: the lake queried without loading it
  • Zero-ETL — from Aurora to Redshift managed by AWS: near-real-time without pipelines
  • Concurrency scaling — automatic extra clusters during dashboard peaks
04 · Numbers and lifecycle

The numbers that matter.

RPU
the serverless unit: starts at 8, billed by the second
3x
the typical speed-up of a proper sortkey on range scans
0 ETL
from Aurora with the zero-ETL integration
PB
the scale of RA3 managed storage
A fast warehouse is data physics: distribution keys, materialized views and WLM — tuned on your real workloads.
05 · Use cases

Where it really pays off.

Reporting direzionale

The balance sheet and sales over years of history: answers in seconds, not batch nights.

Analytics on the ERP

IBM i and ERP data brought to the warehouse: the history that finally speaks.

Datamart consolidation

The ten reporting databases reunited: governance and costs under control.

The warehouse pays off with the right data inside: solid pipelines from the ERP — that's where we make the difference.