Which set of actions should be taken to meet the company’s objectives?
Upload the call recordings to Amazon S3 from the call center. Set up an S3 lifecycle policy to move the call recordings to Amazon S3 Glacier after 90 days. Use an AWS Lambda trigger to transcribe the call recordings with Amazon Transcribe. Use Amazon S3, Amazon API Gateway, and Lambda to host the review and scoring application.
Upload the call recordings to Amazon S3 from the call center. Set up an S3 lifecycle policy to move the call recordings to Amazon S3 Glacier after 90 days. Use an AWS Lambda trigger to transcribe the call recordings with Amazon Mechanical Turk. Use Amazon EC2 instances in an Auto Scaling group behind an Application Load Balancer to host the review and scoring application.
Use Amazon EC2 instances in an Auto Scaling group behind an Application Load Balancer to host the review and scoring application. Upload the call recordings to this application from the call center and store them on an Amazon EFS mount point. Use AWS Backup to archive the call recordings after 90 days. Transcribe the call recordings with Amazon Transcribe.
Upload the call recordings to Amazon S3 from the call center and put the object key in an Amazon SQS queue. Set up an S3 lifecycle policy to move the call recordings to Amazon S3 Glacier after 90 days. Use Amazon EC2 instances in an Auto Scaling group to send the recordings to Amazon Mechanical Turk for transcription. Use the number of objects in the queue as the scaling metric. Use Amazon S3, Amazon API Gateway, and AWS Lambda to host the review and scoring application.
Explanations:
Uploading recordings to S3 and using a lifecycle policy to archive after 90 days to S3 Glacier optimizes storage costs. Amazon Transcribe can be triggered via Lambda for automatic transcription. The review and scoring application can be hosted using S3, API Gateway, and Lambda, providing a serverless and scalable solution. This reduces transcription time and operational overhead.
Using Amazon Mechanical Turk for transcription is not ideal for this use case as it is a manual task-based service. It introduces latency and complexity compared to using Amazon Transcribe, which is fully automated and integrated with AWS services. Additionally, the review application should be serverless or highly scalable, and using EC2 with an Application Load Balancer is overcomplicated and inefficient.
Storing recordings on Amazon EFS is unnecessary and more costly than using S3. AWS Backup for archiving is also redundant as S3 lifecycle policies already handle archiving. Using EC2 instances to process transcriptions is inefficient, and there is no need for additional EC2 resources. Amazon Transcribe is a better solution for transcription.
Sending recordings to Amazon Mechanical Turk for transcription adds unnecessary manual overhead, which is less efficient than using Amazon Transcribe. Using EC2 instances and an Auto Scaling group also adds complexity and cost compared to serverless or simpler solutions like Lambda for transcription. The rest of the architecture introduces unnecessary complexity in managing scaling and transcription tasks.