I think Data Deduplication section should have more detail. I think it is missing discussion about chunk management when updates or modifications are done. I do have a couple of questions about this section. Let's assume that we're keeping chunks of 4K. What happens if a file is modified in the middle due to appending a lot of text (> 1 chunk size)? What would happen in that case? Would the old chunk will replaced by 2 newly created chunks since new data cannot fit in one chunk? Assumption here is due to metadata info, the chunks that contains the data after the modification won't change at all. Or, will the client re-generate the chunks?