DEA-C01 Question 53
Select 3A retail company wants to build an ETL pipeline to process daily sales data from their on-premises database and store it in Amazon Redshift for analytics. The pipeline should ensure data is encrypted during transfer, handle schema changes in the source data, and allow near-real-time processing to update Redshift. Which combination of AWS services and features should you recommend to meet these requirements?
- A
AWS Glue for schema detection and ETL transformation
- B
AWS Database Migration Service (DMS) for near-real-time replication to Amazon Redshift
- C
Amazon S3 as a staging area for incoming data before loading into Amazon Redshift
- D
AWS DataSync for transferring the data from on-premises to Redshift
- E
Amazon Kinesis Data Streams for streaming data directly into Amazon Redshift
Show answer and explanation
Correct answers: A, B, C
Explanation
The correct combination of services ensures the ETL pipeline meets all business requirements. AWS Glue handles schema changes and transformations effectively, AWS DMS provides near-real-time replication with encryption, and Amazon S3 serves as a secure staging area for data before loading into Amazon Redshift. AWS DataSync and Kinesis Data Streams are not appropriate for this specific use case.
- A. Correct.
AWS Glue is ideal for schema detection and ETL tasks, making it a good choice for handling schema changes and transforming the data before loading it into Amazon Redshift.
- B. Correct.
AWS DMS supports near-real-time replication and is well-suited for transferring data securely from on-premises databases to Amazon Redshift.
- C. Correct.
Using Amazon S3 as a staging area is a common pattern for loading data into Amazon Redshift. It provides durability, cost efficiency, and supports encrypted data transfer.
- D. Incorrect.
AWS DataSync is used for large-scale data transfers but is not suitable for near-real-time replication or ETL pipelines involving Amazon Redshift.
- E. Incorrect.
Amazon Kinesis Data Streams is designed for streaming data ingestion; however, it is not directly used for ETL pipelines involving Amazon Redshift. It requires additional components for transformation and storage.