Babelfish: How do I get started?

Learning Resources

• Common links: Product page | Docs | Tutorials/Blogs
• Consider booking an Immersion Day with your Amazon AWS representative.

Giving Babelfish a quick try

Here is a simple tutorial to take a test drive of Babelfish: https://aws.amazon.com/blogs/aws/goodbye-microsoft-sql-server-hello-babelfish/

Important: Always use the latest version of Aurora, 14.3 or higher. Previous versions of Babelfish had minimal SSMS support, making getting started difficult.

You can then connect with SSMS. While not all SSMS feature are supported, you will be able to log in and run queries.

To learn more, visit the AWS Babelfish documentation.

Evaluating a Migration

Use the Compass assessment tool to evaluate the level of support a given app has with Babelfish. Steps:

• Export your app's DDL using SSMS or any other tool.
• Run the Compass tool to generate a report in HTML format.
  • Compass is a Java application, and runs on Windows, Linux or Mac.
  • Example command line: ./BabelfishCompass.sh MyReportName bikestore-objects.sql -replace
• Refactor any code in the DDL that Babelfish doesn't support.
• Run the DDL on a Babelfish-enabled Aurora cluster.
  • Note: No SCT use needed- Babelfish understands the SQL Server dialect.

Migrating Data

• As of Oct 7, 2022 the recommended approach is to use DMS for migration to Babelfish.
• SSIS is another possible tool for importing data into Babelfish. See the tutorial here.
• As time goes on, more data migration tools will be supported. Check back monthly since AWS improves tooling support on a regular basis.

Tips to get Babelfish working & optimized

• You can turn on auto_explain in your Aurora cluster to log how queries are performing. [Tutorial]
• Consider re-optimizing time-sensitive queries to eliminate implicit conversions.
• See a list of what Babelfish does not support here.
• Remember, while Babelfish offers semantically correct execution of T-SQL and the SQL Server dialect, you are running on the Aurora PostgreSQL engine. Therefore, some time-sensitive queries may need to be re-optimized for the Aurora PostgreSQL engine.