Best practices for working with s3 storage?

Question

PHP

The project stored all files in the file system, now it is planned to move to s3, how best to organize file uploads. I don’t understand the moment, if you need to download in a queue, then you must first save to the file system, then read, send and delete the file from the file system? The loads are large and it is not very clear how to organize work with s3.

Answer 1

https://flysystem.thephppleague.com/v2/docs/adapter...

Answer 2

s3 can send files directly to the client, bypassing the server, from the point of view of speed, this is the best option, but it only makes sense if the data is public and does not require authorization (i.e. the maximum can be assumed that the client must log in to receive a link to the object, but if the client recognizes this link once, the second time he can download it without authorization)
you can implement authorization using amazon itself, but this will increase the cost (since you will also need to pay for this layer)
there is an amazon authorization module for the nginx web server, i.e. data can be directly proxyed through its server
ps if the data is needed only by the backend, then by itself load it into memory or disk, depending on the size and need