Digital Reach

186 / 208

While the phone network is highly controlled and generally
static in terms of new countries, Internet networks may come
and go in rapid succession, extend or alter their connectivity,
partition or merge their operations – all of this on a daily basis
– and all without centralized control.
What is a domain name?
For Internet users, the most familiar forms of addresses are e-mail
and Web addresses and the word-based domain names that
appear within them. However, these are quite different from IP
addresses. While humans may prefer names, the machines that
make up the Internet need numbers. Thus, the purpose of the
Domain Naming System (DNS) is to allow Internet services to be
referenced by their domain name.
The DNS comprises thousands of servers across the Internet.
As Figure 1 shows, the DNS operates in a three-step process.
Typically, a computer may query the DNS about a particular name
(for example ; the DNS server will provide the
IP address that corresponds to that domain name (in this
example, 66.102.7.104). From there the computer and the remote
service can communicate directly.
While IP addresses are fundamental to the operation of the
Internet infrastructure, the DNS is, in fact, a service which oper-
ates on the Internet. Although the DNS is an essential service in
today’s Internet, it could be removed or replaced without the need
for changes to the underlying infrastructure. In fact, even if the
entire DNS suffered a catastrophic breakdown, it would still be
possible to connect to any service on the Internet, provided its IP
address is known.
IP addresses management
The Internet started life as an academic and research network,
linking a close community of collaborating organizations and
institutions. Although the early pioneers had no way of predict-
ing the speed and spread of their creation’s growth, they created
an environment that allowed for outrageous success.
A fundamental principle of Internet design is that it is a layered
network that is ‘dumb’ at its core. Network A does not need to
know how Network B is configured in order to talk to it; and
whether it connects by copper wire, optical fibre, radio waves, or
any other infrastructure is irrelevant. So long as a network is
capable of transmitting data according to the standard protocols
(such as TCP/IP) it can become part of the Internet. It is not the
equipment that defines the Internet, but the protocols by which
communication takes place. A key aspect of those protocols is
the system of Internet addressing. Therefore, stable management
of the address system is vital to the success of the Internet.
The early model of address management
Recognising the importance of the Internet’s addressing system,
its founders established a central registry from the outset. The
most basic function of an address registry is to ensure unique-
ness of addresses, so that clashes cannot occur. Originally, this
registry was just one man – the late Jon Postel – using manual
processes to keep track of the addresses allocated to the then
small number of participating networks. The address registry
function became known as the Internet Assigned Numbers
Authority (IANA).
It is easy to forget just how quickly the Internet has grown in
the past decade, but even throughout the 1980s the available
addressing pool seemed vast. The registration function was
straightforward and the allocation mechanisms remained rela-
tively informal. However, as more networks joined the Internet
and growth began to accelerate, it became increasingly clear that
the original address management function would not scale.
The volume of requests, the linguistic and cultural diversity
of requestors, and the complexity of differing regional needs
caused a re-evaluation of the central registry model. In 1992,
the Internet Activities Board of the IETF considered these factors
and concluded: “…it is desirable to consider delegating the
registration function to an organization in each of [the]
geographic areas.” (RFC 1366, 1992)
Another major problem with early address management was
that the address technology was based on a crude subdivision of
the IPv4 address space into three network “classes”, allowing
only three potential address allocation sizes, namely: “Class C”
(256 addresses), “Class B” (65,534), or “Class A” (16,777,214).
By 1993, a new system had been introduced to eliminate the
wasteful practices of “classful” addressing, but even so it was clear
that the address pool could no longer be regarded as limitless.
[ ] 186
DNS
My computer
?
1
68.102.7.104
2
68.102.7.104
3
The
Internet
Figure 1: By a three-step process, DNS provides access to a web site,
using a domain name
Figure 2: Address resources are managed by the five RIRs, under
policies developed by their respective regional communities
Registry Region
AfriNIC Africa
APNIC Asia Pacific
ARIN Canada, United States, several islands in the
Caribbean Sea & North Atlantic Ocean
LACNIC Latin America, Caribbean
RIPE NCC Europe, the Middle East, parts of Central Asia