Updated September 21, 2020
Everyone can be a Star (March 24, 2020)
Examples (November 2, 2019)
Questions and Answers (June 2, 2020)
For Developers (November 2, 2019)

Multiplexed Internet Domain Names

Why and How


Occam's razor - multiplexed domain names are the fastest, easiest, cheapest, most equitable and most judicious way to solve complex domain name problems and provide any individual or organization with a short, appropriate, easy to remember domain name under their preferred domain top level.

Generic name users want .com domains; national users prefer country code domains.  Domain multiplexing can be applied equally to any top level.

Müller is the most common surname in Germany, but there can be only one müller.de
Martin is the most common surname in France, but there can be only one martin.fr
Silva is the most common surname in Brazil, but there can be only one silva.br
Ford is a dictionary term (noun and verb), a place name, a first name, a family name, a company name and a trademark term with several different owners. 
- But there can be only one ford.com

Do you see the problem in this pattern?

Most names, and even trademark terms, are not unique!

Multiplexed domain names introduce a hierarchy within second level names.  This extra level supports multiple use of 'the same' name without ambiguity under the same top-level domain.  Examples could include name.com, name*1.com, name*2.com, etc.

We suggest the asterisk as a multiplexing symbol since it is universally known and often means 'wildcard' (potentially 'one of many'). 

Inserting the asterisk in domain names requires a small code addition.  The asterisk, by design, is part of the character set that cannot be translated by the Internationalized Domain Name (IDN) software that lives in every web browser. 

An IDN-based solution is quick and simple, but wouldn't work in the US and isn't universal.   Universality - the same character meaning the same thing regardless of TLD - is highly desirable for security and interoperability. 

National or regional variants could facilitate the rapid application of name multiplexing however  The Euro sign € provides a simple multiplexing token for members of the  European Union: examples could include müller€1.de, martin€2.fr,  and so on.  The UK could apply the pound sign £. 

Other countries or regions could standardize their own convention. 

Chinese may have a character that means 'one of many' or something similar.  All that's required is agreement within CNNIC (the China Internet Network Information Center) to give 1.4 billion users their own domains.

Improvement could (we think should) be applied equally to all domains, but is needed most in the legacy .com and country code TLDs.

Problems solved by multiplexed names include:

- Anyone can get any name they have a right to, under any TLD. This respects the reality that most names, and even trademark terms, aren't globally unique.

- Users who want a presence on the Internet can own their own identities instead of becoming a product sold to advertisers by social network sites.  Compare the total number of registered second level names with the global population of Facebook users.

- No one can buy and warehouse a domain name to prevent it from being used, or to extract an unreasonable price.  Name speculation, which raises prices by restricting access to names, becomes less profitable.

- The large number of parked names should decrease (and become generally available).

- Domain names would return to their originally intended use as addresses.

- 'Premium' domains would be determined by the value of their content, not by the character string of their name.

- The confusing proliferation of additional ngTLDs could be avoided.

- While it makes most name ownership disputes superfluous, names you 'have a right to' may be restricted depending on jurisdiction or location (country code or geographic TLDs) or type of business.  Examples include .bank, which was always intended to be restrictive, as well as .brand TLDs.

- The technology is transparent and without negative impact on the familiar, existing system.  Prototype multiplexed names have been tested live on the Internet and work within the current domain name system.

- They provide a platform for further development, opening new opportunities.

- They level the Internet playing field for everyone.

- Multiplexed names improve the Internet by opening it for additional 'normal' users.

Modern browsers that combine the address line and a search function already support the concept of multiplexed names in principle. 

Would this proposal grant Verisign, who run the .com registry, an unfair advantage?  No, since this proposal treats all top-level domains equally, user preference is the final arbiter.

If you're concerned about Verisign's position and believe the Internet, including domain names, to be a human right, then consider how ICANN or the ITU, the US or the UN, could claim eminent domain and require mandatory auctions to run the .com registry, or run it themselves, or reallocate responsibility by some other means.  Any proceeds above the cost of operation could help finance the relevant authority's pro-Internet activities. 

Remember however that Verisign invested $21 billion to buy Network Solutions (NSI) in 2000, and with it the right to run the .com registry.  They took a significant business risk to achieve their position.

Domain names as we know them were introduced in 1983.  The only major improvement since then was the introduction of Internationalized Domain Names that support characters and scripts outside standard ASCII English.  Internationalized Domain Names were introduced in 2003!

ICANN's new generic Top Level Domains are not an improvement, they don't support normal users.  When the US Department of Commerce created ICANN their primary concern was:
 "... widespread dissatisfaction about the absence of competition in domain name registration."

Commerce never mentioned customer benefit or user demand.  ICANN was built on supply side push - Internet users were never asked if they wanted thousands of new Top Level Domains.  The results show they didn't.

The DoC also runs the US Patent and Trademark Office.  They could have said: "compete by inventing a better system" but instead ICANN was founded and populated by groups that supported launching new TLDs.

More than 1900 ngTLD applications were submitted and over 1200 new TLDs have been delegated (made active) since October, 2013, but they don't solve user problems.

The legacy .com TLD continues to grow.  The number of registered .com domains is over 149 million (http://research.domaintools.com/statistics/tld-counts/).

The 266 county code TLDs together add at least another 160 million registrations.

How have ICANN's 1,200+ new generic top-level domains worked out?  Not very well.

- The total number of ngTLD registrations peaked below 30 million in April, 2017, and then fell below that number for years.  Growth from November, 2019, pushed the count past 32 million (https://namestat.org/s/newgtld-summary) but the curve has gone negative since early March to 30.6 million.

- The ngTLDs have lost 306,000 registrations in the past 4 weeks (source: namestat.org/s/newgtld-summary).  Legacy .com added 231,000 in the same period.

- The largest ngTLD registers 5.6 million second level names (well below earlier numbers); only 8 of 1238 ngTLDs register more than 1 million names.

- On September 1, the top 10 ngTLDs account for 62 of all ngTLD registrations (https://ntldstats.com), and the top five alone account for 46.5%!  The bottom 42% of ngTLDs register 10 domains or fewer!  

- New generic TLDs can gain or lose thousands of registrations in a day - gains are based on intensive marketing and, we can presume, domainer purchases for intended resale at inflated prices.  Losses are a result of non-renewals.  It's a game, not a value proposition for users.

- Failing to see value materialize, numerous organizations have cut their losses and withdrawn their .brand ngTLD applications.

Unfortunately many domain names, under both legacy and new generic top levels, aren't actually used.  

If ngTLDs don't resolve domain name problems, what does ICANN propose to do?  They're still talking about opening yet another round of ngTLD applications!

Isn't it better to expand the Internet by providing unlimited names instead of the ngTLDs merchants want to impose?

The Internet Domain Name System was never designed to provide universal naming, and that caused a number of problems now so ingrained that most people accepted them as inevitable.  Things as common as domain name disputes, name warehousing and auctions, and the drive to market unwanted new TLDs are consequences of the system not being designed to provide universal naming.

The basic problem: most people, companies and trademark holders can't use their own name under their preferred top-level domain.

The DNS is a technical system written to a technical specification.  Problems can be resolved by extending the specification.  This was done when Internationalized Domain Names made it possible to use 'foreign' characters and scripts in domain names.

Since the domain name system is hierarchical, multiplexed names add a hierarchical level within second level names.  It's like adding street numbers to street names within a city.  This makes the name-space under any top-level domain virtually unlimited.

We need a new token character to identify/generate a hierarchy.  We suggest the asterisk as a multiplexing token, together with a number or letter(s).

Compare it to the accepted character that designates email addresses.  MaratSade.fr could be seen as a domain name, but write it Mar@Sade.fr and you recognize it instantly as an email address.  The same transparency can apply to domain names if * is the character indicating multiple use of the same name.

The asterisk is often used as a wild card character indicating 'one of many'.  The final number (or letter) indicates 'which one' of several users of the same name.

Technically, names are registered in a set format and translated by a simple edge application to include the multiplexing token. This is similar to, but not the same as, the system used to generate foreign characters and scripts in Internationalized Domain Names.

Aren't domain names ICANN's responsibility?  Can't you just add numbers, or edit the software already used to translate foreign characters?
See the answer to those and other relevant questions under: Questions and Answers  (Updated June 2, 2020).

When did it become appropriate to let organizations like ICANN and the IETF define the problems, create the solutions, and then universally sanction those solutions without competition or independent oversight?  Do you want Boeing to design, manufacture, and approve its own aircraft without competition or FAA review? 

Multiplexed domain names are a technical step toward Universalized Domain Names.  Multiplexed names nested with Internationalized Names make Universalized names available to anyone, anywhere, in any language or script - under their country code, legacy, or preferred new generic top-level domain.

Is Internet access a human right?  In that case registering your own name, under your preferred TLD, should also be your right.  Otherwise you support a privileged minority owning their names and excluding the majority.

Registering your own domain name isn't a problem when Everyone can be a Star. (Updated March 24, 2020)

Multiplexed names are not offered as a supported product; we want to demonstrate how the Internet domain name system can evolve to eliminate unnecessary restrictions and provide relevant names for everyone.  We have no interest, past or present, in any domain registry, registrar, or re-seller.

The combined URL address line/search field in modern browsers is ready to support multiplexed domain names through disambiguation. 

Multiplexed domain names under our test follow all applicable Internet standards, but the translation format used in our tests is not standardized.

Last updated September 21, 2020

W. Kenneth Ryan