A large-scale data lake on S3 contains billions of small objects encrypted with AWS KMS. The security team is concerned about the high cost of KMS API calls and the potential for hitting KMS request rate limits. Which solution should they implement to reduce costs and increase performance?
S3 Bucket Keys are designed specifically for this scenario. By decreasing the frequency of requests from S3 to KMS, they reduce the cost of encryption by up to 99% and help avoid throttling. This is highly effective for data lakes where a massive number of objects are frequently accessed or uploaded.
Why this answer
S3 Bucket Keys reduce the number of requests to AWS KMS by generating a bucket-level key that is used to derive unique data keys for objects, thereby lowering KMS API call costs and avoiding request rate limits. This is the recommended solution for large-scale data lakes with many small objects encrypted with SSE-KMS.
Exam trap
SCS-C03 often tests the misconception that switching to AWS Managed Keys reduces costs or increases limits, but the correct solution is S3 Bucket Keys to minimize KMS calls.
How to eliminate wrong answers
Option A is wrong because AWS Managed Keys (aws/s3) do not have higher request limits or lower costs compared to Customer Managed Keys; they are also not recommended for fine-grained control. Option C is wrong because CloudFront does not cache KMS metadata; it caches content, and 'Origin Request Caching' for KMS metadata is not a feature. Option D is wrong because re-encrypting with SSE-S3 removes KMS-based encryption, which may violate security requirements; it also does not reduce KMS costs but eliminates them, which is not the goal if KMS is required.