AWS SSA-C003 #21

An ecommerce company wants to launch a one-deal-a-day website on AWS. Each day will feature exactly one product on sale for a period of 24 hours. The company wants to be able to handle millions of requests each hour with millisecond latency during peak hours.
Which solution will meet these requirements with the LEAST operational overhead?

A. Use Amazon S3 to host the full website in different S3 buckets. Add Amazon CloudFront distributions. Set the S3 buckets as origins for the distributions. Store the order data in Amazon S3.
B. Deploy the full website on Amazon EC2 instances that run in Auto Scaling groups across multiple Availability Zones. Add an Application Load Balancer (ALB) to distribute the website traffic. Add another ALB for the backend APIs. Store the data in Amazon RDS for MySQL.
C. Migrate the full application to run in containers. Host the containers on Amazon Elastic Kubernetes Service (Amazon EKS). Use the Kubernetes Cluster Autoscaler to increase and decrease the number of pods to process bursts in traffic. Store the data in Amazon RDS for MySQL.
D. Use an Amazon S3 bucket to host the website’s static content. Deploy an Amazon CloudFront distribution. Set the S3 bucket as the origin. Use Amazon API Gateway and AWS Lambda functions for the backend APIs. Store the data in Amazon DynamoDB.

---

The best solution for this scenario would be Option D.

This option involves using an Amazon S3 bucket to host the website’s static content, deploying an Amazon CloudFront distribution with the S3 bucket set as the origin, using Amazon API Gateway and AWS Lambda functions for the backend APIs, and storing the data in Amazon DynamoDB.

This solution meets the requirements of handling millions of requests each hour with millisecond latency during peak hours, and it also minimizes operational overhead. AWS Lambda can handle the burst in traffic, and DynamoDB is designed to handle high request rates and provide consistent, single-digit millisecond latency.

Moreover, this solution is serverless, which means there’s no need to provision, maintain, or administer servers, and you only pay for what you use. This makes it a cost-effective choice as well.

Please note that while the other options could potentially work, they may require more operational overhead and might not be able to handle the high request rate with low latency as effectively as Option D. For example, Option A does not provide a database service for order data, which is crucial for an e-commerce application. Options B and C involve managing servers or containers, which increases operational overhead.