Build a borderless Lakehouse

Learn how to build a borderless Lakehouse that unifies data silos across Amazon Web Services (AWS), Google Cloud, and AlloyDB for PostgreSQL without the need for complex ETL.

You use Lakehouse for Apache Iceberg as the central intelligence hub, AlloyDB as an operational data source, and Managed Service for Apache Spark for high-performance vectorized processing. Finally, you use Gemini to derive powerful business insights from your lakehouse.

Scenario

In this scenario, you have data across the following sources:

  • Transactional data (users, orders, order items) in an operational AlloyDB database.
  • Product data in an AWS S3 bucket.
  • Clickstream event logs in Cloud Storage.

You join these datasets to identify target demographics for your next marketing campaign and generate personalized outreach emails.

What you learn

In this codelab, you perform the following tasks:

  • Set up and configure Lakehouse catalog endpoints to unify data silos across AWS S3, Cloud Storage, and AlloyDB.
  • Run high-performance vectorized query processing using Managed Service for Apache Spark.
  • Join transactional, product, and clickstream datasets without data movement or complex ETL pipelines.
  • Use Gemini to derive business insights and generate personalized marketing outreach from your lakehouse.

Take the codelab

Launch the step-by-step codelab:

Start the borderless Lakehouse codelab

What's next