Why is versioning required for cross-region replication of AWS S3 buckets?

Question

This is just a purely theoretical question. For CRR Amazon requires Versioning on the source and destination buckets. Is this just to reduce their costs in terms of re-replicating the same object?

Answer 1

AWS documentation mentions that CRR is built on top of S3’s existing versioning facility.

My guess is, as the replication is asynchronous, it would need to 'non-changing' copy of data during replication, and versioning makes it easy. So while the replication is in-progress, same object/file can still be modified (new copy upload or delete etc).