What should the solutions architect do to accomplish this?
Provide an API hosted on an Amazon EC2 instance. The EC2 instance performs the required computations when the API request is made.
Design a REST API using Amazon API Gateway that accepts the item names. API Gateway passes item names to AWS Lambda for tax computations.
Create an Application Load Balancer that has two Amazon EC2 instances behind it. The EC2 instances will compute the tax on the received item names.
Design a REST API using Amazon API Gateway that connects with an API hosted on an Amazon EC2 instance. API Gateway accepts and passes the item names to the EC2 instance for tax computations.
Explanations:
Hosting on a single EC2 instance does not provide scalability or elasticity, especially during peak times when response times are critical.
API Gateway combined with AWS Lambda is scalable and elastic, allowing the system to handle increased traffic automatically by scaling out Lambda functions as needed. This solution is cost-effective and can manage fluctuations in traffic efficiently.
Using an Application Load Balancer with two EC2 instances provides limited scalability and may not be sufficient during high-demand periods. It also requires more maintenance than a serverless option like Lambda.
API Gateway and an EC2 backend lack elasticity as EC2 instances do not scale as seamlessly as Lambda functions, and this setup can lead to increased latency under heavy load.
As I understand it, the answer is:
Design a REST API using Amazon API Gateway that accepts the item names. API Gateway passes item names to AWS Lambda for tax computations.