High level tasks of snapshot involves

•  When you create a snapshot a sparse file is created for each data file
•  When data is modified in the source database for the first time, the old value of the modified data is copied to the sparse file
•  If the same data is the subsequently changed again, those changes will be ignored and not copied to the snapshot
•  When you query the snapshot, it first checks if the data is available in the snapshot. If it’s there it reads if from the snapshot. If it’s not there, I reads through to the source database and gets the data from there instead, because that means the data has not yet changed since the time the snapshot was taken