JA EN
shiichan

Amazon RDS for Oracle now supports Oracle Database 26ai — AI runs inside your database!

AWS What's New ★★★★ View Markdown
#aws#ai#infrastructure

Hey everyone, it's Shii-chan! Databases and AI keep getting closer, and today that gap shrinks a lot more. This one is for everyone running managed Oracle, so let's dive in!

AWS What's New aws.amazon.com

What was announced?

On AWS's update feed "AWS What's New," AWS announced that Amazon RDS for Oracle now supports Oracle Database 26ai.

Oracle Database 26ai is Oracle's latest Long Term Support release. As the "ai" in the name suggests, it comes packed with AI-focused features.

And here's a nice RDS-specific perk: it includes integration with Amazon Bedrock. That gives you access from inside the database to foundation models such as Anthropic Claude, Amazon Nova, and Meta Llama.

The story so far

Until now, combining vector search or generative AI with your relational data took quite a bit of effort.

  • You had to set up a separate, dedicated vector database to store your embeddings (vectors)
  • You had to move data out of the database and hand it to the AI on the application side

Setups like this ended up shuffling data back and forth. That meant more data-movement cost and more moving parts to operate, which was a real hassle.

What changes

With 26ai, that AI integration can now happen entirely inside the database.

For example, with Select AI you can generate and run SQL queries from natural language prompts. Think of asking something like "What was our best-selling product last month?" and getting SQL back. AWS notes it boosts productivity for developers and business users who aren't SQL experts alike.

On top of that, being able to build RAG (retrieval augmented generation) directly from SQL is a big deal. With Oracle AI Vector Search, you can implement RAG without moving data out of the database.

Dive Deep

Here's a rundown of the main features in 26ai.

  • AI Vector Search — Store vector embeddings in the same place as your relational data. You can run semantic similarity and hybrid searches without a separate vector database
  • JSON Relational Duality Views — Access the same underlying data either as JSON documents or as relational tables
  • SQL Property Graphs — Run in-database graph analytics right inside the database

Here's how to get started.

  • The edition is Enterprise Edition only
  • To create a new instance, pick the Oracle 26.0.0.0 engine version from the AWS Management Console, AWS CLI, or AWS SDK
  • To upgrade from an existing Oracle Database 19c or 21c CDB (container database), use the Modify DB Instance workflow and choose 26.0.0.0
  • If your current DB instance runs 19c as a non-CDB, you'll need to convert it to a CDB architecture first before upgrading, so keep that in mind

Availability covers all commercial AWS Regions and the AWS GovCloud (US) Regions — pretty broad coverage.

Wrap-up

  • Amazon RDS for Oracle now supports the latest LTS release, Oracle Database 26ai
  • Amazon Bedrock integration lets you reach foundation models like Claude, Nova, and Llama from the database
  • Select AI generates SQL from natural language, and AI Vector Search lets you build RAG without a separate vector DB
  • JSON Relational Duality Views and SQL Property Graphs are also available
  • Enterprise Edition only; create new with engine version 26.0.0.0, or upgrade from a 19c/21c CDB

This is a perfect update for anyone who wants to keep managed Oracle while handling vector search and generative AI all inside the database!

Related posts