S3 IAM Account vs HTTP Get

I need some advice. I have data hosted in AWS S3 that client software downloads as needed.

I originally wrote it so it downloaded it via a public http get command. But then - reading more, I thought I needed to use read-only group IAM credentials. (Custom credentials for each user would be too onerous I think).

I ran some performance tests and I find that http Get is usually faster (10-25%) than the AWS-SDK, and JClouds is consistently much slower.

What are the advantages to using a Read-only IAM account vs http get? I think they ultimately provide the same access. I thought the SDK would be faster, but that is not what I find.