Category: Tech matters
There is no single name system that is necessarily bound to the Internet. Unlike IP addresses, which are in every IP packet, names are an application construct, and, in theory, applications have considerable latitude in how they handle such names. There could be many name systems that could coexist within the Internet — in theory.
In practice, there is strong peer pressure to use a single name system. For a network to be useful to its users we need to use the network to communicate references to accessible resources and services, and we would like to do so within a human-oriented name space. We’d like to name such resources in some fashion and then pass these names to each other. And we’d like to assume that the name in my context refers to the same digital resource as it will in your context. For this to work, we need to draw these names from a common name space and resolve these names into network coordinates in a coherent and consistent manner. Such name coherence within the Internet is in everybody’s interest and market disciplines become the most effective way to ensure that the name space remains unified and cohesive.
For this reason, we tend to assume there is a single Internet name space, and that space is defined by the Domain Name System (DNS).