Data Retrieval and Data Transfer OUT in Internet

Question

Hi,

I have 32TB of data stored on S3 Standard bucket and would like to make a copy to my on-premise SQL Server (local machine) via SQL Server Integration Services (SSIS).

(1) I am wondering does it count as Data Retrieval or Data Transfer OUT in Internet? 
(2) If I use aws s3 sync (aws CLI API) to download the data to my hard drive, does it count as Data Retrieval or Data Transfer OUT in Internet? 
(3) The rate ($0.09/GB) for Data Retrieval or Data Transfer OUT in Internet is the same for S3 Standard plan, I am wondering would it help to reduce the cost (based on my need) if I switch to S3 Standard-IA and perform data retrieval actions?

Wish my questions can be answered. Thank you very much.

Accepted Answer

The best option with the 32 TB of data is to directly copy from S3 standard to your on-prem server. This will be charged as Data transfer Out. The reason being the data is transferred from S3 to your network through internet and  retrievals charges are not applicable.

Please note , S3 transfer out are based on tiered pricing with $.09/GB is for up to 10TB and is lower after that.

If you switch to S3 Standard -IA- 0.01$/GB Retrieval fee applies to put the data from IA back into S3-Standard. AWS then applies ,The $0.09 data transfer fee from S3 to outside .In addition to that you will pay for the GET request as well ($.01 per 10,000 requests)

Answer

All of the scenarios you mention will be counted as Data Transfer Out - the data is being transferred from S3 to your network across the internet. This is regardless of the storage class in S3 and the tools used to retrieve that data.

You can reduce the cost by using Direct Connect - the data transfer charges are lower when you're using a private network to do the transfer. However, for one-time transfer purposes it's not really practical.

Data Retrieval and Data Transfer OUT in Internet

Relevant content