Objects in S3 can not be simply mapped to AOS objects since updating is handled differently. In AOS an update always creates a new object with a new unique ID, while in S3 the objects ID (the key) would remain the same and only a new revision (version) is created.
A mapping could look like this:
Each new object has a unique key as a separate field equivalent to the key in S3 to identify and object across all of its revisions that can be set by the user.
In the S3 API only this key will be referenced, although internally the id is used.
This would probably require changes in the database and API.
Objects in S3 can not be simply mapped to AOS objects since updating is handled differently. In AOS an update always creates a new object with a new unique ID, while in S3 the objects ID (the key) would remain the same and only a new revision (version) is created. A mapping could look like this: