I've been able to easily duplicate this behavior on Server 2012 R2 and Server 2008 utilizing different PPTP VPN endpoints. The problem is not with the PPTP connection itself but what it does to the rest of the network stack. I noticed when setting up a PPTP connection to a vendor site (and unchecking to use the remote gateway) from one of our terminal servers, that users who accessed apps from our terminal server complained of lengthy logins and windows switching. As soon as I click disconnect on the PPTP connection, performance is immediately as expected. A real quick and easy way to see this issue happen is to run netstat -b; when not connected to a PPTP connection it flies through the list as it should but when connected to a PPTP connection, it crawls, taking 1-4 seconds to print each line entry.

I have tried multiple differences all with the same scenario:

1. Disabling IPv6
2. Not setting any automatic routes on PPTP connection
3. Using remote default gateway
4. Different Server OS (2008)
5. Different VPN endpoint (HideMyAss)

At this point it seems "built-in" or by design; would like some clarification on what I'm seeing and ultimately am looking for a way to have this PPTP live permanently connected on this terminal server, without affecting application/network stack performance.

Hi TyroneS2S,

When VPN is connected, it would consume system resource to encrypt decrypt network packets. It may cause the performance decrease.

Besides, VPN client would use the DNS server configured on VPN server. It may cause the delay when loging on and browsing webpage.

About the command netstat -b, according to the description of the -b flag, this option can be time-consuming. I suppose that connecting VPN adds a active connection and caused the command to take more time to execute.

Best Regards,

Leo