Cheers for the info! I guess that's where I got the wrong idea from - attempting to stress test one machine from another machine, I'd always hit that limit, but now I understand why.