Geographic vs. Traditional Domain Hierarchies: A Look at Internet Naming Schemes

When the Domain Name System (DNS) was formed in the mid-80's, several top-level domains (TLDs) were created to contain hosts in the new system. The "ARPA" TLD was a transitional domain used during the migration to DNS, the five categorical TLDs "GOV", "EDU", "COM", "MIL", and "ORG" were used for the bulk of the name space, and the two-letter ISO3166 country codes were reserved for geographical TLDs [1]. The Internet was still primarily a US academic and research network, and the vast majority of new domains were registered in the categorical TLDs.

As Internet began to grow internationally, many countries implemented domain hierarchies of their own based on the geographic country code TLDs, despite most of the traditional, categorical TLDs being available to all. In the United States, most hosts continued to use the traditional top-level domains, instead of the geographical "US" domain [2].

Today, most new domain registrations in the United States are within the traditional TLDs, while new domains within other countries tend to be within geographic TLDs. In recent years, policy changes have forced many entities, such as schools and local governments, to seek geographic domain names [3].

The traditional top-level domains provide Internet sites with simple addresses composed of their name and the TLD, such as "NAME.COM". This is beneficial in that users can have short and easy to remember mail addresses, and URLs are short enough to be related in conversation or displayed in television commercials. If the entity moves to a different geographic location, there is no need to change domain names.

However, there are a number of problems with the traditional TLDs. The structure of these TLDs is very flat, with a potentially enormous number of domains under each TLD. The zones for these TLDs are growing massive. At the time of this writing, the zone for the "COM" TLD is almost 23MB. With the expected increase in usage over the next few years as the Internet becomes an all-encompassing medium, these zones could grow to an unmanageable size, straining the precious root nameservers. It will also become increasingly more difficult to obtain a meaningful domain name as most pronounceable utterances are registered.

Another consequence of the flat domain space is the concentrated administrative authority. When InterNIC decided to start charging for domain registrations in September 1995, although probably justified due to the huge administrative task of handling the enormous flat domain space, many people were disgruntled and could do nothing but accept the decision.

With geographical domains, the name space is fanned out among many zones in a more hierarchical fashion than the traditional domains. With entities registering domains in their country's geographical name space, they don't have to compete for names internationally. The load and administration of geographical domains is distributed, so no one set of nameservers takes on the bulk of the burden, and policy decisions are made locally.

One disadvantage of geographic domains is the often long and more complex names, making them hard to remember and convey. For instance, large countries, such as the United States, usually break up the geographic top-level domain by locality, which lengthens the names. This may not be much of an issue, though, in the coming days where most users will use domain names in the form of URLs embedded in hypertext links, safely hidden from view. Long names can be minimized, though, by cities using common abbreviations instead of the full name. For instance, a site in Minneapolis might have the name "name.minneapolis.mn.us", while a site in San Francisco might have the much more preferable name of "name.sf.ca.us".

When selecting a domain name for an entity connecting to the Internet, a geographical name should at least be considered. Many internet service providers don't inform their customers that a geographical name is available. Some entities, such as K12 schools in the US, only fit under a geographical domain. Choosing a geographical name gives a site a more community-oriented presence, is Internet-friendly, and avoids the starting and recurring InterNIC fees for a traditional name.

In the coming years, the traditional top-level domains will probably have to be closed to new registrations or changed significantly due to the immense increase in zone sizes and the lack of unique names. Migration to geographical domains will probably happen in time without intervention anyway, because the overcrowding in the traditional TLDs will begin to turn people away. In the United States, nationwide entities cannot feasibly register a geographic name under the present system, which may lead to more lenient options in the US domain, such as a "COM.US".

As society allows individuals to become more mobile, many users that are always on the move may outgrow geographical names. Notebook computers, pagers, and even wrist watches will all be on Internet, and a hierarchy based on geography may not suffice. Some discussion of a new hierarchy for individuals has taken place, but no reasonable method of organizing and fanning out such a hierarchy has been proposed. Chances are, a hierarchy flexible enough to handle all 5 billion people on the planet would involve names no more meaningful than the telephone numbers or social security numbers of today.

References

[1] Postel, J., and J. Reynolds, "Domain Requirements", RFC-920, USC Information Sciences Institute, October 1984.

[2] Cooper, A., and J. Postel, "The US Domain", RFC-1480, USC Information Sciences Institute, June 1993.

[3] Postel, J., "Domain Name System Structure and Delegation", RFC-1591, USC Information Sciences Institute, March 1994.


Last modified: December 1995
David Simmons
send mail