The options at a glance
- Amazon RDS for SQL Server: SQL Server as a managed service. AWS takes care of the installation, patches, backups and Multi-AZ failover. You choose the version: SQL Server 2016 to 2025.
- SQL Server on Amazon EC2: your own SQL Server on a virtual machine, with full control and all features, but you manage it yourself. RDS Custom for SQL Server sits in between.
- Babelfish for Aurora PostgreSQL: Aurora PostgreSQL with a T-SQL layer. Applications keep using SQL Server drivers and T-SQL, while the data lives in PostgreSQL.
Amazon Aurora itself does not run SQL Server: Aurora is compatible with MySQL and PostgreSQL. When people mention "SQL Server to Aurora", they usually mean Babelfish, or a full conversion to Aurora PostgreSQL.
Amazon RDS for SQL Server: what you lose
RDS keeps your database on SQL Server, so most code keeps working unchanged. SQL Server Agent, Database Mail, cross-database queries, distributed transactions, SSIS and SSRS are available. What you lose is access to the server itself: there is no sysadmin role, no access to the operating system and no file system.
- CLR assemblies, FILESTREAM and
xp_cmdshellare not available - Server triggers are not available, and linked servers and Service Broker only with restrictions
BACKUPandRESTOREin your own code: use the RDS stored procedures for native backup and restore via Amazon S3- Windows authentication works through AWS Managed Microsoft AD
- Server settings are managed through DB parameter groups instead of
sp_configure
Babelfish: T-SQL on PostgreSQL
Babelfish lets Aurora PostgreSQL accept connections over the SQL Server protocol and run T-SQL. That can remove SQL Server licence costs without rewriting the whole application at once. Babelfish versions are tied to Aurora PostgreSQL versions: Babelfish 3 runs on Aurora PostgreSQL 15, up to Babelfish 6 on Aurora PostgreSQL 18.
Babelfish does not support everything SQL Server can do. SQL Server Agent, CLR, Service Broker, Database Mail, distributed transactions and in-memory tables are not available; queries across databases work only in a limited way. Every statement Babelfish does not support has to be rewritten. AWS provides Babelfish Compass to produce a compatibility report on your scripts.
Migration methods
- To RDS for SQL Server: native backup and restore via Amazon S3, or AWS Database Migration Service (DMS) with a full load followed by change data capture to keep downtime short.
- To Babelfish: generate the structure scripts with SSMS, run them against the Babelfish endpoint, and move the data with AWS DMS.
As with every move, logins, Agent jobs, linked servers and anything outside the database need their own plan. For storage, RDS for SQL Server goes up to 256 TiB with additional storage volumes; an Aurora cluster volume up to 256 TiB on recent versions.
How to choose
If the application has to keep working with as few changes as possible, RDS for SQL Server or EC2 is the safe choice. If licence costs are the main reason to move and the code uses few SQL Server-specific features, Babelfish is worth a serious look. If the code relies on CLR, file access or xp_cmdshell, EC2 or a redesign of that part is unavoidable.
How the free scan helps
The scan reads the structure of your database with a read-only script and scores it against Amazon RDS for SQL Server, Babelfish, Aurora PostgreSQL and the other targets. You see which objects use features that are not available on each, so you can compare the routes on facts. The analysis runs in your browser and nothing is uploaded.
Questions
Can Amazon Aurora run SQL Server?
No. Aurora is compatible with MySQL and PostgreSQL. Babelfish for Aurora PostgreSQL lets it understand T-SQL and the SQL Server protocol, within the features Babelfish supports.
Which SQL Server versions does Amazon RDS offer?
Amazon RDS for SQL Server offers SQL Server 2016, 2017, 2019, 2022 and 2025. AWS announces end dates per version, so check them when you choose.
Do I get sysadmin rights on RDS?
No. The master user has broad rights, but not sysadmin, and there is no access to the operating system. Tasks such as native backups run through stored procedures supplied by AWS.
Can Babelfish replace SQL Server for any application?
Not for every application. It works best for code that uses common T-SQL. Features such as SQL Server Agent, CLR and Service Broker need an alternative.
See how your database scores on Amazon RDS for SQL Server, Babelfish and Aurora PostgreSQL: run the read-only script and paste the result. Free, and nothing leaves your browser.
Start the free scan