In order to answer this I'll lean on an answer I modified to another question. A good example of a URI is how you identify an Amazon S3 resource. Let's take:
s3://www-example-com/index.html
[fig. 1]
which I created as a cached copy of
http://www.example.com/index.html
[fig. 2]
in Amazon's S3-US-West-2 datacenter.
Even if StackOverflow would allow me to hyperlink to the s3://
protocol scheme, it wouldn't do you any good in locating the resource. Because it Identifies a Resource, fig. 1 is a valid URI. It is also a valid URN, because Amazon requires that the bucket (their term for the authority
portion of the URI) be unique across datacenters. It is helpful in locating it, but it does not indicate the datacenter. Therefore it does not work as a URL.
So, how do URI, URL, and URN differ in this case?
NOTE: RFC 3986 defines URIs as scheme://authority/path?query#fragment