This actually came up on one of the internal aliases. So I thought it would make a good tip.
Unlike SIS (Single Instance Store), Windows Deduplication does not have a central database to track redundant data that has been consolidated. Windows Deduplication uses reparse points within each deduplicated file that points to the consolidated data within the ‘chunk store’.
SIS requires that two or more files be absolutely the same before file system redundancies can be removed. However, Windows Deduplication only needs for ‘parts’ files to be the same.