tprc’s 29th research conference on communication information cabbages and

kings. october 2001 

dr. milton l. mueller

graduate program in telecommunications and network management

syracuse university school of information studies 

abstract

the internet domain name system (dns) is a hierarchical name space that enables the

assignment of unique mnemonic identifiers to internet hosts and the consistent mapping

of these names to ip addresses. the root of the domain name system is the top of the

hierarchy and is currently managed by a quasi-private centralized regulatory authority

the internet corporation for assigned names and numbers (icann). this paper

identifies and discusses the economic and policy issues raised by competing dns roots.

the paper provides a precise definition of root-competition and shows that multiple roots

are a species of standards competition in which network externalities play a major role.

the paper performs a structural analysis of the different forms that competing dns roots

can take and their effects on end-user compatibility. it then explores the policy

implications of the various forms of competition.

the thesis of the paper is that root competition is caused by a severe disjunction between

the demand for and supply of top-level domain names. icann has authorized a tiny

number of new top-level domains (7) and subjected their operators to excruciatingly slow

and expensive contractual negotiations. the growth of alternate dns roots is an attempt

to bypass that bottleneck. the paper arrives at the policy conclusion that competition

among dns roots should be permitted and is a healthy outlet for inefficiency or abuses of

power by the dominant root administrator. 

1. introduction and overview 

three years after the creation of the internet corporation for assigned names and

numbers (icann) management of the domain name system (dns) continues to

generate controversy. one of the most intense and significant policy controversies

associated with icann is the problem of competing dns roots. within internet circles

the problem of competing roots has taken on the characteristics of a religious war.

internet “catholics” a collection of veteran techies acknowledging vint cerf as their

pope and the late jon postel as the messiah demand allegiance to the one true

universal root. [iab 2000; lynn 2001] internet protestants with elected icann

board member karl auerbach assuming the role of luther insist on the freedom to

create alternate roots – but just as luther’s challenge led to the proliferation of hundreds

of sects and denominations so (some fear) will the proliferation of dns roots fragment

the internet.

my goal in this paper is to identify and discuss the economic and policy issues raised by

competing dns roots. i begin with the observation that multiple roots are a species of

standards competition in which network externalities play a huge role. i then perform a

structural analysis of the different forms competition among dns roots can take and

analyze their compatibility effects and policy implications.

in the established domain name market the value that can be added by competing dns

roots seems to be small relative to the compatibility and fragmentation risks.1

 so why are

we getting competing dns roots? the answer is clear: root competition is caused by

icann’s extremely restrictive and slow addition of new top-level domains to the domain

name system. there are no technical constraints preventing the addition of thousands of

new top-level names to the dns root. there are many willing suppliers of new top-level

domains and many members of the public are willing to pay to register them. the

icann regime however has deliberately maintained an extreme scarcity in the supply

of top-level domain names. it has authorized a tiny number of new top-level domains (7)

and subjected their operators to excruciatingly slow and expensive contractual

negotiations. the growth of alternate dns roots is an attempt to bypass that bottleneck.

more fundamentally competing roots are about competing sources of authority over the

internet. the basic issue they raise is: should the direction of the dns market be driven

by technology and the market or by a quasi-governmental collective action via icann?

the policy conclusion i draw is that competition among dns roots should be permitted

and is a healthy outlet for inefficiency or abuses of power by the dominant root. such

competition does not threaten the universality of the internet interconnection becauseboth suppliers and end users have powerful economic incentives to remain compatible

and connected with each other. indeed the deck is stacked so heavily in favor of an

established root that if an alternate root achieves critical mass and threatens the

dominance of the existing root it can only happen because the existing root is doing

something seriously wrong. icann itself claims to be based on an “internet community

consensus.” if so it has nothing to fear from competing roots because – given the

importance of network externalities in this market – a sustainable alternate root could

only arise from a clear lack of consensus support for icann policies among a critical

mass of suppliers and consumers.

a more difficult and complex issue to be faced is how much interconnection should be

implemented between a dominant root and alternate systems. if icann continues to

stifle the market for new top-level domains this issue will have to be faced. in this paper

i sketch out some of the issues involved in interconnecting roots but do not make any

detailed or specific proposals.