Something I've always wondered, but never bothered enough to look into ...
Say I have 2 Windows machines on a network. Let's be very creative and call them WIN1 and WIN2. And let's use the $ admin shares.
On WIN1, I have drive S: mapped to \\WIN2\C$. I also have drive T: mapped to \\WIN2\D$
Using Windows Explorer on WIN1, I copy a file from S: to T:. What happens:
1) The file is read from WIN2 over the network to WIN1, and written back over the network to WIN2.
2) Some magic kicks in and WIN2 writes the file directly to the destination UNC path (i.e. it does go through WIN2's NIC) and Windows somehow simulates permissions.
3) Some AI magic kicks in and WIN2 writes the file directly to its D: drive (i.e. it bypasses its own NIC.)
4) some combination of the above based on what Windows decides is possible and most efficient
1) is massively inefficient, but probably most likely (consider permissions, if S: and T: are on different machines / networks, etc)
2) would be way more efficient, but not super efficient
3) would be pie-in-the-sky amazing.
Say I have 2 Windows machines on a network. Let's be very creative and call them WIN1 and WIN2. And let's use the $ admin shares.
On WIN1, I have drive S: mapped to \\WIN2\C$. I also have drive T: mapped to \\WIN2\D$
Using Windows Explorer on WIN1, I copy a file from S: to T:. What happens:
1) The file is read from WIN2 over the network to WIN1, and written back over the network to WIN2.
2) Some magic kicks in and WIN2 writes the file directly to the destination UNC path (i.e. it does go through WIN2's NIC) and Windows somehow simulates permissions.
3) Some AI magic kicks in and WIN2 writes the file directly to its D: drive (i.e. it bypasses its own NIC.)
4) some combination of the above based on what Windows decides is possible and most efficient
1) is massively inefficient, but probably most likely (consider permissions, if S: and T: are on different machines / networks, etc)
2) would be way more efficient, but not super efficient
3) would be pie-in-the-sky amazing.

