How can a file sync service avoid storing the same bytes twice?

Answer

Split files into chunks and name each chunk by the hash of its content. An identical chunk, from another user or another version of the file, has the same name and is stored once; the client checks which hashes the server lacks and uploads only those. The cost: tracking references before deleting a chunk.

Review this in your daily deck → All cards in Storage