I’m getting why more bandwidth is needed for upload - for redundancy purposes.
But why downloading 110 MB file uses 1002 MB of bandwidth?
It looks possible to download chunk from single node and only in case if it is corrupted, download it from other node.