2개 답변
- 최신
- 최다 투표
- 가장 많은 댓글
0
A glue ETL job can do that for you .
- Read the DMS data
- Write the data back to a different S3 bucket based on your partition keys .
- Loop it for each table through a Glue Python job
On the high level its a 2 step process , minimal coding , use DynamicFrames which will also help you in clubbing multiple CDC files into larger chunks improving better reads on the consuming applications .
Please note
- Glue will incur costs
- Its not going to be real time as DMS , since additional read and writes are involved
- DMS will result in multiple versions of the record , For example , Inserts , Multiple Updates and Delete . You can handle that in Glue, but the program will have a few more steps then. This is based on your requirement though
답변함 2년 전