AWS SCT
Downloadable tool that converts database schema and code from one engine to another before the data moves.
Verified against AWS documentation · 2026-09-09
Key facts
- 01SCT converts the structure — tables, views, stored procedures, functions, triggers. DMS moves the rows
- 02Only needed for heterogeneous migrations (Oracle → Aurora PostgreSQL). Same engine both sides needs no conversion
- 03Assessment report grades the job up front and marks exactly what won't convert automatically
- 04Reality check: AWS now points you to DMS Schema Conversion (managed, in-console) first — SCT is the download for paths it doesn't cover
Exam triggers
Phrases and scenarios that point here.
- "Different database engine", "heterogeneous migration", "Oracle to PostgreSQL" → SCT + DMS
- Stored procedures, views, or triggers must come across → SCT, not DMS alone
- "Estimate the migration effort before committing" → SCT assessment report
- Same engine on both sides (MySQL → RDS MySQL) → DMS alone, SCT is the wrong answer
Not this service
- Moving the actual data → AWS DMS
- Migrating whole servers or VMs, not databases → AWS Application Migration Service
Use cases
- 01A team sitting on Oracle with 900 PL/SQL stored procedures wants Aurora PostgreSQL — SCT converts what it can and hands back a list of what a human has to rewrite
- 02Scoping a migration before anyone promises a date: run the assessment report to turn "how hard is this?" into a count of objects and hours
- 03Teradata or Netezza warehouse → Redshift, with SCT data extraction agents pulling data locally and staging it through S3 (or onto a Snowball device when the pipe is too slow)
- 04Escaping commercial database licensing — convert SQL Server to Aurora PostgreSQL so the license line disappears from the bill
- 05Converting embedded SQL inside Java, C#, or C++ application code so the app doesn't break the day the database underneath it changes
How it's charged
- 01The tool itself is free — download it, run unlimited conversions, no AWS charge for SCT
- 02You pay for everything it touches: the target RDS/Aurora/Redshift, the DMS replication instance moving data, S3 storage for extracted data
- 03Large schemas are usually converted from an EC2 instance sitting near the source database — that instance is billed normally
- 04Data extraction agents for warehouse migrations run on your own hosts or EC2; that compute is yours to pay for
- 05DMS Schema Conversion, the managed alternative, is also free to use — you pay only for the S3 storage it consumes