Using CloudFront, is it still useful to use different S3 region?

Question

Up to now, we were using S3 to store our files, using buckets in different regions to be closest to our data generator and people getting data (much more GET than POST, POST'er typically closer to GET'er).

We are moving to CloudFront for many reasons. So now the data is pushed and got from the closest CloudFront endpoint from the user, as a proxy to/from S3.

The question that now arises is whether it is still useful for any reason to store our data on a bucket depending on the region?

• GET will not be faster as they are served from the CF endpoint, except for the very first GET of an CF area after a "long" duration without GET
• POST will not be faster as they are pushed to the CF endpoint
• The cost of CF does not seem to be affected by the region of the origin S3

Answer 1

As you said, region may not make any significant difference for GETs, since Amazon CloudFront distributions have a single endpoint: cloudfront.amazonaws.com

However if you are writing (put) to S3 directly from your user side, here it might be better to keep it at a closer region.