A customer-facing application has a relational data model and needs frequent complex queries (joins and aggregations), but it also experiences a significant read-heavy workload. Which design choice best improves read performance while keeping relational features?

Use DynamoDB with a single partition key and avoid indexes to keep writes simple.

DynamoDB is great for key-value and document patterns, but complex joins/aggregations require redesign. Avoiding indexes would likely hurt query flexibility and performance. This option conflicts with the relational query requirement.

Store the data in S3 and query it directly from the application without a database.

S3 is object storage and is not designed for low-latency relational query workloads. Querying directly would be slow and operationally complex. Complex joins/aggregations would not be straightforward.

Switch the database to DynamoDB but keep using the same relational SQL queries and joins.

DynamoDB does not support SQL joins in the same way relational databases do. Moving to DynamoDB generally requires a query-pattern redesign with keys and indexes. Keeping the same query approach would not work as intended.

What does this SAA-C03 question test?

Static NAT maps one inside address to one outside address.

What is the correct answer to this question?

The correct answer is: Add read replicas to an RDS or Aurora cluster and keep the primary for writes. — For read-heavy workloads on a relational schema, read replicas are the most direct performance optimization that preserves relational functionality. With RDS or Aurora, you can replicate the database and direct read queries to replicas while writes remain on the primary. This reduces load on the writer, improves read throughput, and helps maintain performance during traffic spikes. DynamoDB would require redesigning queries and data access patterns because joins/aggregations are not supported the same way as in SQL databases. Why others are wrong: Option A pushes the problem into DynamoDB without addressing the need for complex relational queries; avoiding indexes further limits performance and query flexibility. Option C removes the database layer entirely and would not meet low-latency relational query requirements. Option D assumes DynamoDB can use the same relational SQL with joins, which is not how DynamoDB works. Read replicas best match the stated relational + read-heavy needs.

What should I do if I get this SAA-C03 question wrong?

Then try more questions from the same exam bank and focus on understanding why the wrong options are tempting.