I was wondering if anyone had done any checking to see if buffering works faster than non-buffered remote views?

My secenario is that I am updating a SQL server database via a modem, and it seemed to me that if I buffered my views, it would write all the changes when I issued tableupdate(). The delete commands take forever to process and the insert commands are slightly longer.

I'm working with live data, so a test is kinda scary to perform.

Thanks!