I wonder if it is something with permissions in the smart switch, gonna put an unmanaged gigabyte switch in later and see
As I mentioned in the other thread, we have had to run down sporadic issues with NewCrop, and in the process, we were encouraged to use fiddler2 ( a web-proxy/testing tool) to observe the traffic in the background between AC and NewCrop. I'll buy a round if the variability isn't right there.
Some of our infrastructure happens to be in the same dark Tier-4 co-location facility, so we get 10-20 millisecond response (small world), but that makes the response time from their systems that much more noticeable.
Checking local network is always essential, but sometimes it IS the other guy.
