A retail company wants to implement a system on AWS that processes customer product reviews. The system must detect the sentiment of the reviews, translate non-English reviews to English, and save the results. The team wants to deploy this solution with the lowest possible operational overhead, avoiding tasks like operating system patching, capacity provisioning, and software installation. Which of the following designs best adheres to the AWS Cloud design principle of 'services not servers' to accomplish this goal?
- Ingesting reviews via Amazon API Gateway, processing them using AWS Lambda functions that call Amazon Comprehend and Amazon Translate, and storing the final data in Amazon DynamoDB.Cevap
- BDeploying a single, high-capacity Amazon EC2 instance that runs a web server, a custom natural language processing library, a translation database, and a local PostgreSQL database.
- CDeploying the application as a monolithic service across a fleet of Amazon EC2 instances managed by an Auto Scaling group, using Amazon CloudWatch metrics to scale the instances in and out based on processing demand.
- DProvisioning dedicated Amazon EC2 instances pre-configured with AWS-managed machine learning software, and configuring AWS Systems Manager to automate the operating system patching on these instances.
Cevap
Ingesting reviews via Amazon API Gateway, processing them using AWS Lambda functions that call Amazon Comprehend and Amazon Translate, and storing the final data in Amazon DynamoDB.
The correct design uses Amazon API Gateway, AWS Lambda, Amazon Comprehend, Amazon Translate, and Amazon DynamoDB. Each of these is a serverless or fully managed service. With this architecture, AWS handles server provisioning, operating system patching, scaling, and high availability, which completely removes the operational burden of managing servers from the customer.
Adım Adım Çözüm
Anahtar Kavram
The AWS Cloud design principle of 'services not servers' guides architects to use fully managed and serverless services to minimize operational overhead and focus on application logic rather than infrastructure maintenance.