I'm going to try and sum up how DCP is intended to be used. But first, a quick overview of what you need to use it
Your application (for you, the website running on the webserver) uses a caching system; it is a memory area where the application puts everything it feels like calculating only once. Usually, things that are very time consuming are put in cache.
When you have multiple web servers, each server uses its own cache; but since every server is supposed to handle the same website and the samùe data, each server will compute the same information for its personnal use.
The distributed cache is supposed to handle this problem; very roughly, it is a process running on each machine that needs to work with shared information; the distributed cache gives the machine access to data shared between all the machines. When a machine puts something in the cache, every machine sees it.
DCP acts as the bridge between your application and the distributed cache; the application doesn't know what type of cache its working with, but hands DCP the data to save. DCP then uses the configuration you've set up to determine what to do with the information, and interact with the distributed cache.
So to sum up, you do need a distributed caching cluster installed on the machines you intend to use DCP on. Setting up and configuring these cache clusters is not explained in DCP's documentation since there are many options available, but since we support AppFabric (the distributed cache solution by Microsoft) and Memcached (an open source distributed cache solution) you can try and read on these two tools to find out which one you can use.