By design, communication between the server you operate locally and S3 is routed via the public internet. The same channel of communication is used for the servers in AWS as well. Even though S3 is an AWS service, and when you access S3 from your EC2, by default, it will get routed via public endpoints and through the open internet.
Suppose, you store a software installation on S3, and you want to install the software on your on-premise server locally. From your on-premise server, you can use an AWS CLI, such as PowerShell, to download the software from S3. You specify the name of the bucket and object in your AWS account with a cmdlet. This communication is done via HTTP over the public internet. Now, suppose you want to do the same thing on your AWS servers. In this case as well, even though you are getting the software from S3, it will still route...