I'm looking at a couple of systems and their resultant perfmon logs. They are averaging 4 to 10 times the number of datagram received discarded packets than my baseline numbers, and as many as 23,372 'datagram received errors' in a 2 hr period when the expected level is... zero. The description at least for the 'datagram received discarded' value, it points that the 'lack of buffer space' could be the cause. But doesn't remind one where to check said nbuffer space / variable. Any ideas / suggestions on where I should be checking next?

I can't actually test this solution yet as internal policy is stopping me from actually having access to the boxes I'm troubleshooting / benchmarking. But I've found an exerpt which states that if I add a value to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\NetBT\Parameters\ of Value Name: MaxDgramBuffering Data Type: REG_DWORD Value: 0x40000 That it will / should double the effective cache available to datagrams. (unless this value is already configured / the datagram cache has already been increased to 256KB in some other manner) Mucking with the Registry to 'test' stuff isn't usually the safest. I would really like to hear some input back to confirm what I'm intended to do / propose to test. Thanks, -Aaron