>Hi Al,
>
>I have tried it both ways Client & Server same machine vs Client & Server different machines. I get the same result.
>
>It is interesting that the delay is not with SQL pass-thru and once the remove view is created the performance is fine. It is just in the initial creation of the remote view that results in the delays. The initial creation of any remote view involves slight delay. I see that with MS-SQL as well, but no where near as significant: Seconds vs Minutes!

Is Vista SP1 applied? My understanding is it fixed a number of file system-related performance issues.

If real-time antivirus scanning is enabled, can you try disabling it for testing?

Is the remote view definition in a .DBC or are you generating it on the fly? If in a .DBC, could you try locating that in a test folder in the root off the root of C: to see if that makes a difference? My understanding is Vista "virtualizes" certain folders by default, if your .DBC is in one of those it might be worth testing with it elsewhere.

This is a bit of a long shot, but it might be worth testing with SMBv2 disabled on the Vista machine:

http://blogs.technet.com/askperf/archive/2008/05/30/two-minute-drill-overview-of-smb-2-0.aspx