Question

I need to upload a large number of large-size objects from different Geographic locations to an S3 bucket. What is the best mechanism to do so in a fast & reliable way?

A. I can connect to an application running on AWS EC2 that is hosted in multiple regions using Route 53 & use latency based routing to upload files to the S3 bucket.

B. I can use a Direct Connect link from each of the Geographic location for transferring data quickly.

C. I can use S3 Transfer Acceleration from each Geographic location that will route the data from their respective Edge locations to S3.

D. I can directly access the S3 bucket from the different locations & use a multi-part-upload for transferring huge objects.

Answer 1

Answer: C.

Option A is incorrect since Route 53 latency routing only calculates latency between different endpoints based on the Internet traffic & location proximity rather than optimizing the network for fast data transfers.

Option B is incorrect since Direct Connect is used for very specific purposes like extreme security requirements for the data transfer.

Also, establishing multiple Direct Connect infrastructures would be expensive from a cost standpoint.

Option C is CORRECT.

The best way to address this scenario is to route the requests to the nearest CloudFront edge location from the different Geographic locations.

Edge locations provide a fast network infrastructure bypassing much of the internet for delivering content to S3 destinations.

Performance gains of nearly 50 - 500% can be observed while using S3 Transfer Acceleration.

Option D is incorrect.

It is possible to use S3 endpoints directly for data transfer.