Question #989
A company processes video files using a fleet of Amazon EC2 t2.micro instances that pull tasks from an Amazon Simple Queue Service (Amazon SQS) queue. Each task is processed in 10 seconds or less. The instances store the processed videos in an Amazon Elastic File System (Amazon EFS) volume mounted on all instances. Metrics indicate that instances are idle when the SQS queue is empty. A solutions architect must redesign the architecture to optimize costs.
Which combination of steps will meet these requirements MOST cost-effectively? (Choose two.)
Use c5.4xlarge instances instead of t2.micro instances for video processing. Reduce the fleet size by 75%.
Convert the video processing into an AWS Lambda function. Configure the Lambda function to pull tasks from the SQS queue.
Modify the video processing to store results in Amazon DynamoDB.
Modify the video processing to store results in Amazon RDS PostgreSQL with auto-scaling enabled.
Modify the video processing to store results in Amazon S3.
Explanation
The current architecture uses EC2 instances that incur costs even when idle. By converting to AWS Lambda (B), the solution leverages serverless computing, which processes tasks on-demand with no idle costs. Lambda's event-driven model efficiently pulls tasks from SQS, ensuring cost-effectiveness. Storing results in Amazon S3 (E) replaces the expensive EFS storage, as S3 offers lower-cost, scalable storage for large files like videos.
Other options are incorrect:
- A: Larger instances (c5.4xlarge) are costlier than t2.micro, even with a reduced fleet.
- C/D: DynamoDB and RDS are unsuitable for large video files due to item size limits and higher storage costs.
Key points: Use serverless (Lambda) for short, bursty workloads and S3 for cost-effective storage of large files.
Answer
The correct answer is: BE