Q27 — AWS SAA-C03 Ch.5

Question 27 of 65 | ← Chapter 5

Q327. A company wants to host a scalable web application on AWS. The application will be accessed by users from different geographic regions of the world. Application users will be able to download and upload unique data up to gigabytes in size. The development team wants a cost-effective solution to minimize upload and download latency and maximize performance.What should a solutions architect do to accomplish this?

Correct Answer: C. Use Amazon EC2 with Auto Scaling and Amazon CloudFront to host the application.

Explanation

To provide a scalable, high-performance, and cost-effective solution for hosting a web application on AWS, Amazon EC2 with Auto Scaling and Amazon CloudFront can be used. EC2 instances can be deployed across multiple regions of the world to minimize latency, and Auto Scaling can be used to automatically scale the infrastructure up or down to meet the changing number of users accessing the application.Amazon CloudFront can be used to improve performance by caching frequently accessed content at edge locations located nearest to the application users. This reduces the round-trip time required to request and receive data from the origin server and minimizes the download latency. Using Amazon S3 with Transfer Acceleration (Option A) can also be an option, but since the application involves uploading and downloading unique data up to gigabytes in size, it may not be a suitable solution as it would require uploading all data to S3 before using it via the application and could lead to increased latency.Using Amazon S3 with CacheControl headers (Option B) is not recommended since CacheControl headers work well when there is frequent reuse of objects, which is unlikely in the case of a web application where users upload and access unique data.Using Amazon EC2 with Auto Scaling and Amazon ElastiCache (Option D) can be used to provide scalable, high-performance infrastructure, but it does not address the issue of minimizing upload and download latency.