Among the many improvements in Windows Vista is an entirely rewritten TCP/IP stack. The new stack support some very cool stuff such a PNRP, IPv6, Teredo, improved QoS support, and a new congestion control protocol called Compound TCP, or CTCP for short.
CTCP promises to greatly increase the efficiency of high latency connections. Most of today's congestion protocols (which determine how fast to send data over the network) leave a great deal to be desired when it comes to dealing with high latency connections. They end up under utilizing the available pipe. What does that mean?
It means that if you have a connection that spans a long distance, like from Boston to Los Angeles, even if that connection has a "fat pipe" (it can support very large transfer rates), you'll often get significantly less than the maximum transfer rate. So while you should be getting 8Mb/s, you might actually only get 5Mb/s. The reason for this is that, for most congestion protocols, high latency is a signal to slow down because something is wrong, when in fact it's just the nature of the connection.
CTCP handles things differently. Honestly, the math is way over my head... but the end result is something we can all comprehend: MORE POWER... er... speed... um... faster downloads.
Since I'm running Windows Vista on my trusty IBM T42P laptop, I decided to kick the tires on this fancy new TCP stack. The first thing I had to do was enable CTCP, as it comes disabled by default. To do this you must enter the command shown in the screen shot to the right. Be sure to do this in a command prompt running as Administrator.
The exact command is: netsh interface tcp set global congestion=ctcp
Once you've done this you should be enjoying your new super-fast net connection, right? Well, not so much. The testing that I've done shows almost no change in my transfer rates. I conducted my tests using Speakeasy's nifty speed test page. These numbers are the average of 5 tests each for both my transfer rates from Boston to NYC, and from Boston to LA.
As you can see, while I gained a bit during downloads from NYC, I lost a bit during uploads. Strangely enough, the much higher latency connection to LA seemed to only suffer from turning on CTCP.
So what's the conclusion? Well, I really doubt I created the proper environment to showcase CTCP. I know Microsoft's tests show that it can more than double the bandwidth available on certain connections. I'm also less than confident that using a flash speed test app was probably not the most scientific way to test this. Furthermore, a Comcast cable connection (8Mb/768Kb) is also probably not the best way to judge CTCP's effectiveness.
So my conclusion is that I'll defer to the judgment of people who are more qualified to test this than I am.