For the "examples" section...
I've had many problems with an Outlook 2003 client at home which connects over a VPN to Exchange (ancient) at my office, there is an ADSL connection at home and a new connection at work (serviced office, shared internet connection).

Recently the Outlook client would get into a "freeze" state for varying lengths of time (with that annoying balloon popping up all the time), sometimes being unusable for an hour or more. Over those periods I could connect a Remote Desktop client to the same server, albeit a bit lumpy in its response.

Tests with command line ping didn't show anything except a bit of extra packet loss, the latency didn't look that much worse.

The Outlook connection was RPC (not over HTTPS), from my reading I gathered that this is susceptible to latency.

Well, from my chart I would say it (or the way it uses the VPN) is more susceptible to packet loss; I now have a nice graphic that shows clear 10 to 30% packet loss coinciding with a couple of hours of absent connection (PingPlotter is running on the Exchange server, ping target is a reliable site, early hops show packet loss on parts of route used by the VPN).

Thanks! OK, I'll buy it now.


Attachments
2040-2008-02-08 08-00 whole day see clear pkt loss at 18-00 (Sanitised).png