A recent discussion on CF-Talk had a member trying to find instructions for setting up ColdFusion in "Distributed mode" on a CF 11 server. If you have never heard of "distributed mode" you are not alone among CF developers. It's not a common setup in my experience. Here's a quick 101. Many processes and daemons on a given server communicate using the TCP stack. TCP provides a predictable, high capacity, mature protocol for piping requests and data in and out of server processes. In this case, IIS or Apache communicate with CF through the local loopback (localhost or 127.0.0.1) IP address and a port - usually 8012 (or 8013 or 8014) chosen at installation. Naturally you can alter the IP address to which you are connecting, changing it from the loopback to... well any IP that's listening on 8012. That means you can set up your ColdFusion servers separately from your web servers. After all it's just IP networking. Why would you do that? The Muse will let his guest handle that question.
Meanwhile, to preempt (or perhaps spur) discussion, the Muse will note that this process is similar to something called "Reverse proxy" which functions in much the same way. The difference being with CF distributed, IIS on the front end handles all the "http stuff" and passes the request to CF just like it was a local engine, whereas with reverse proxy the HTTP request is simply redirected to the alternate server. That means the alternate server needs to be a full webserver plus application server. While that increases the overhead a bit, it has some advantages - but that's a topic for a different post.
Back to our CF-Talk question, it quickly became apparent that not many folks actually knew how to accomplish this task on a CF 10 (or 11) server due to the underlying platform switch to Tomcat. After some back and forth Byron Mann chimed in with some very specific instructions on how to get this done. Byron is a lead engineer at HostMySite - which makes the Muse feel better about their ColdFusion support. :) Here's his tutorial.