Well, I just wondered if IBM was clever enough to start transmitting the
object used in the SAVRST* family of commands to store the object(s) being
transmitted before the save was even complete, then why not start the
restore before the whole encapsulated object is received?

After all, you and I cannot send a save file before the save is done so
how can IBM?
put /qsys.lib/rob.lib/rob.file /qsys.lib/rob.lib/rob.file
229 Entering Extended Passive Mode (|||21589|).
150 Sending file to member ROB in file ROB in library ROB.
426 Data transfer ended.
226 ABOR subcommand successful.
Therefore if IBM can get past this block, why can't they get past the
restore block? Therefore speeding up restore and reducing the disk space
used on the target machine?

Let me guess, the logic is that the save happens much faster than you can
fathom the transmission occurring so it makes sense to start the send
while the save is still going on. However, the restore should occur much
faster then the communication occurs therefore the restore would be
waiting and waiting for more data?
Wonder if that would be true for a system with a single 35GB drive and a
10GB ethernet connection?

Rob Berendt




Return to Archive home page | Return to MIDRANGE.COM home page