Domain Name System
The Domain Name System: A Non-Technical
Explanation – Why Universal Resolvability Is Important
What is the Domain Name System?
The Domain Name System (DNS) helps users
to find their way around the Internet. Every computer on the Internet
has a unique address – just like a telephone number – which is a
rather complicated string of numbers. It is called its "IP address"
(IP stands for "Internet Protocol").
But it is hard to remember everyone's IP
address. The DNS makes it easier by allowing a familiar string of
letters (the "domain name") to be used instead of the arcane IP
address. So instead of typing 188.8.131.52, you can type
www.icann.org. It is a "mnemonic"
device that makes addresses easier to remember.
Translating the name into the IP address
is called "resolving the domain name." The goal of the DNS is for any
Internet user any place in the world to reach a specific website IP
address by entering its domain name. Domain names are also used for
reaching e-mail addresses and for other Internet applications.
What is universal
resolvability and why is it important to users?
Think of the phone system . . . when you
dial a number, it rings at a particular location because there is a
central numbering plan that ensures that each telephone number is
unique. The DNS works in a similar way. If telephone numbers or domain
names were not globally unique, phone calls or e-mail intended for one
person might go to someone else with the same number or domain name.
Without uniqueness, both systems would be unpredictable and therefore
Ensuring predictable results from any
place on the Internet is called "universal resolvability." It is a
critical design feature of the DNS, one that makes the Internet the
helpful, global resource that it is today. Without it, the same domain
name might map to different Internet locations under different
circumstances, which would only cause confusion.
When you send an e-mail
to your Aunt Sally, do you care who receives it?
Do you care if it goes to your Uncle Juan
instead? Wait a minute…do you have an Uncle Juan? Then whose Uncle
Juan received it? Do you care if it reaches Aunt Sally if you send it
from work but my Uncle Juan if you send it from home?
Of course you care who receives it . . .
that's why you wrote it in the first place. Whether you're doing
business or sending personal correspondence, you want to be certain
that your message gets to the intended addressee.
If at any point the DNS must make a choice
between two identical domain names with different IP addresses, the
DNS would not function. It would not know how to resolve the domain
name. When a DNS computer queries another computer and asks, "are you
the intended recipient of this message?", "yes" and "no" are
acceptable answers, but "maybe" is not.
Where does ICANN come
This is where ICANN comes in . . . ICANN
is responsible for managing and coordinating the DNS to ensure
ICANN is the global, non-profit,
private-sector coordinating body acting in the public interest. ICANN
ensures that the DNS continues to function effectively – by overseeing
the distribution of unique numeric IP addresses and domain names.
Among its other responsibilities, ICANN oversees the processes and
systems that ensure that each domain name maps to the correct IP
What goes on behind the
Behind the scenes, the story becomes a
little more complicated.
In an Internet address – such as icann.org
– the .org part is known as a Top Level Domain, or TLD. So-called "TLD
registry" organizations house online databases that contain
information about the domain names in that TLD. The .org registry
database, for example, contains the Internet whereabouts – or IP
address – of icann.org. So in trying to find the Internet address of
icann.org your computer must first find the .org registry database.
At the heart of the DNS are 13 special
computers, called root servers. They are coordinated by ICANN and are
distributed around the world. All 13 contain the same vital
information – this is to spread the workload and back each other up.
Why are these
root servers so important?
The root servers contain the IP
addresses of all the TLD registries – both the global registries such
as .com, .org, etc. and the 244 country-specific registries such as .fr
(France), .cn (China), etc. This is critical information. If the
information is not 100% correct or if it is ambiguous, it might not be
possible to locate a key registry on the Internet. In DNS parlance,
the information must be unique and authentic. Let us look at how this
information is used.
Scattered across the Internet are
thousands of computers – called "Domain Name Resolvers" or just plain
"resolvers" - that routinely cache the information they receive from
queries to the root servers. These resolvers are located strategically
with Internet Service Providers (ISPs) or institutional networks. They
are used to respond to a user's request to resolve a domain name –
that is, to find the corresponding IP address.
So what happens to a
user's request to reach our familiar friend at icann.org?
The request is forwarded to a local
resolver. The resolver splits the request into its component parts. It
knows where to find the .org registry – remember, it had copied that
information from a root server beforehand – so it forwards the request
over to the .org registry to find the IP address of icann.org. This
answer is forwarded back to the user's computer. And we're done. It's
that simple! The domain name icann.org has been "resolved"!
Why do we need
the resolvers? Why not use the root
After all, they contain essentially the
same information. The answer is for reasons of performance. The root
servers could not handle hundreds of billions of requests a day! It
would slow users down.
If you are still with the story, you are
already wondering about more complicated names with more parts such as
www.icann.org. Well, the DNS is a
hierarchical system. First, the resolver finds the IP address for the
.org registry, queries that registry to find the IP address for
icann.org, then queries a local computer at that address to find the
final IP address for www.icann.org.
Just what you would expect.
It is important to remember the central
and critical role played by the root servers that store information
about the unique, authoritative root. Confusion would result if there
were two TLDs with the same name: which one did the user intend? The
beauty of the Internet architecture is that it ensures there is a
unique, authoritative root, so that there is no chance of ambiguity.
What about "alternate
roots?" How do they fit into this picture?
Anyone can create a root system similar to
the unique authoritative root managed by ICANN. Many people and
entities have. Some of these are purely private (inside a single
corporation, for example) and are insulated from having any effect on
the DNS. Some, however, overlap the authoritative global DNS root by
incorporating the unique, authoritative root information, and then
adding new pseudo-TLDs that have not resulted from the
consensus-driven process by which official new TLDs are created
through ICANN. The alternate root operators persuade some users to
have their resolvers "point" to their alternate root instead of the
authoritative root. Others (New.net is a recent example) also create
browser plug-ins and other software workarounds to accomplish similar
effects. The one uniform fact about all these efforts is that these
pseudo-TLDs are not included in the authoritative root managed by
ICANN and, thus, are not resolvable by the vast majority of Internet
Why do alternate roots
create a problem?
There are many potential problems caused
by these unofficial, alternate root efforts to exploit the stability
and reach of the authoritative root. These efforts are often promoted
by those unwilling to abide by the consensus policies established by
the Internet community, policies designed to ensure the continued
stability and utility of the DNS.
First, the names of some of these pseudo-TLDs
could overlap TLD names in the authoritative root or those that appear
in other alternate roots. Our familiar friend icann.org could appear
in two different roots. Your e-mail to Aunt Sally could end up with my
Second, the unknowing users might not be linked to one of these
alternate roots and not be able to reach these pseudo-TLD addresses at
all. Your e-mail to Aunt Sally could end up as a dead-letter.
Third, those purchasing domain names in these pseudo-TLDs may not be
aware of these and other consequences of the lack of universal
resolvability. Or they may be under the impression that they are
experiencing universal resolvability when in fact they are not. They
may be very upset to learn that the names they registered are also
being used by others, or that a new TLD in the authoritative root will
not include those names.
These problems are not significant so long as these alternate roots
remain very small, that is, house few domain names with little
potential for conflict. But if they should ever attract many users,
the problems would become much more serious, and could affect the
stability and reliability of the DNS itself. Users would lose
confidence in the utility of the Internet.
What is ICANN's role?
ICANN's mission is to protect and preserve
the stability, integrity and utility – on behalf of the global
Internet community – of the DNS and the authoritative root ICANN was
established to manage. ICANN has no role to play with alternate roots
so long as these and other analogous efforts do not create
instabilities in the DNS or otherwise impair the stability of the
authoritative root. But ICANN does have a role to play in educating
and informing about threats to the Internet's reliability and
ICANN is a consensus development body for
the global Internet community, and its focus is the development of
consensus policies relating to the single authoritative root and the
DNS. These policies include those that allow the orderly introduction
of new TLDs.
There are those–including operators of
commercialized alternate roots–who pursue unilateral actions outside
the ICANN consensus-development process. Many hope to circumvent these
processes by claiming to establish some prior right to a top-level
domain name. ICANN, however, recognizes no such prior claim. ICANN
will continue to reflect the public policy consensus of the global
Internet community over the private claims of the few who try to
bypass this consensus.
In Short . . . . . .
Just as there is a
single root for telephone numbers internationally, there must be a
single authoritative root for the Internet, administered in the public