
Accelerating your Database Migrations Using Gen/AI - Lets Talk About Data
Join us to see how AI makes database migrations easier and faster! We'll show you practical ways to use AI tools that help convert database code, map data structures, and spot potential issues early. Whether you're moving to the cloud or upgrading your database, learn how AI can do the heavy lifting and save you time. Perfect for anyone involved in database projects who wants to work smarter, not harder.
Tony Mullen
Amazon Employee
Published Feb 4, 2025
The show discusses how DMS has evolved from supporting just a few key relational databases at launch to now handling around twenty different sources and targets, including NoSQL databases and data warehouses.
A significant portion of the discussion centered around the new Generative AI capabilities in DMS Schema Conversion, launched at re:Invent. The team explained how this feature helps convert database schemas and code that previously couldn't be automatically converted using traditional rule-based methods. They emphasized that the Gen AI functionality is free, optional, and includes security measures like data obfuscation before processing. Through a live demo, Mike showed how the Gen AI feature could convert complex Oracle database functions to PostgreSQL, highlighting both its capabilities and limitations.
The speakers stressed that while Gen AI may not always provide perfect conversions, it significantly accelerates the migration process by providing a starting point that requires minimal manual fixes. They also discussed broader migration challenges beyond technical conversion, including business buy-in, testing, security compliance, and the importance of proper planning for successful database migrations.
Key Highlights:
- DMS Schema Conversion with Gen AI is free and optional for customers
- Gen AI capability currently supports Oracle/SQL Server to PostgreSQL migrations
- Data is obfuscated before being processed by the AI model
- The service uses Amazon Bedrock with custom RAG implementations
- Migration timelines are often extended by non-technical factors like testing and security compliance
- Gen AI may not provide perfect conversions but significantly reduces manual work
- The service continues to evolve based on customer feedback and usage
- Integration with Q Developer allows for application code conversion
Tony Mullen - Senior Database Specialist @AWS
John Winford - Senior Manager Database Services @AWS
Mike Revitt - Database Migration Specalist @AWS
Guy Baulch - Specialist Solutions Architect
- Re:invent talk - https://www.youtube.com/watch?v=ElzTxmS6XYE
- AWS DMS - https://aws.amazon.com/dms/
- DMS Diagnostic support scripts - https://docs.aws.amazon.com/dms/latest/userguide/CHAP_SupportScripts.Oracle.html
Any opinions in this post are those of the individual author and may not reflect the opinions of AWS.