AWS’s S3 is best advertised as “infinitely scaling storage”. It utilizes buckets (directories) which contain objects (files). These buckets are created at the region level, but the name must be unique across all regions and all accounts. The max object size is 5 TB and if the upload size is greater than 5 GB, then a multi-part download will need to be utilized. There is no real directory system either, each object links to a key, that may look like this s3://my_bucket/myfolder/object. The myfolder name isn’t a directory path exactly it is simply part of the key name that links to that object, you can have multiple objects accessible through that path but there isn’t a “myfolder” folder on s3 that contains these objects, you can browse S3 on your AWS management console as if they are folders and it will automatically delete all objects that contain /myfolder in their name if you delete /myfolder. S3 itself is a global service despite the buckets being created per region. By default block-all public access is enabled for your S3 bucket, when you open your S3 object, such as a txt file from AWS you are given a pre-signed URL that authenticates you and allows you to view the object, there is also a public URL but by default if anyone tries to access the object that way they’ll get a access denied failure. Your objects are safe with AWS S3!