It is quite clear from the previous article, decreasing the http requests has an practical limit. We can not decrease http requests to the level that one server can handle parallel at a time; we designated the number as n and also said that the value of n is usually 5.
So, if we can distribute the components, take that this website has (2n-1) number of http requests; then the first half; that the up to the nth (not the thing we generally denote by using nth; it is actually the 5th here) requests will load at a time and the rest; that is (n-1) will have to wait still the previous requests are downloaded AND a latent period will be there too.
So, if we can distribute our content across different servers that will reduce the loading time. CDN is a collection of geographically distributed multiple servers to distribute content more efficiently to the user:
There are various companies those offers the service of Content Delivery Network including Amazon CloudFront, Rackspace, Microsoft Cloud services and many others.
There is a fantastic reference text on CDN published by Springer Edited by Rajkumar Buyya, Mukaddim Pathan and Athena Vakali ; for advanced users or those who want to deploy CDN for serious works, we would recommend to read it.
For WordPress, we can easily configure if we use W3 Total Cache or W3 Super Cache.
The article we wrote is even less than a primer; many difficult things are made easier by not using various terminologies. It is advisable that you consult a professional for the optimum result.