REVIEW: Development of the Domain Name System

From: Nilton Bila <nilton_REMOVE_THIS_FROM_EMAIL_FIRST_at_cs.toronto.edu>
Date: Mon, 3 Oct 2005 02:10:00 -0400

REVIEW: Development of the Domain Name System

The paper examines the design of Domain Name System and reports on the
experiences collected during the initial stages of its implementation
between 1983 to 1988, pointing out its successes as well as shortcomings.
Before going into the design objectives, it motivates the need for the DNS
as by 1982 it was becoming apparent that its predecessor, the HOSTS.TXT
system, was unable to cope with the growing size of the internet since its
file size grew proportionately and it was costly to send it to hosts. The
DNS was thus conceived to, with litle overhead, fully replace the
HOSTS.TXT system while providing scalability, performance and
interoperability among existing protocols, at the time.

The implementation of the DNS has been extremely successful as its
original design left room for additions and provided a lot of flexibility.
By internally structuring its name space as variable depth tree, it
provides support for different semantics, it supports networks other than
the DARPA Internet (by means of classes), it allows for additions such as
the MX record type for mail servers, it divides its namespace into zones
which allows organizations to have authority over their domains
simplifying the task of adding hosts within an organization. The authors
were visionary in seeing a system that would scale well with the
exponential growth in the number internet hosts. The paper is good at not
only glorifying the design of the DNS but also pointing out its
shortcomings, such as the need for a methodology to design and implement
new classes and types, and the unforeseen need for negative caching.

Although among its design goals was to not impose any obvious size
limitations, the implementation of DNS system did impose a 63 octect limit
per label (associated with a node) and a total of 263 octects per name,
the paper fails to point the defficiency as such, however.

It is the visionary goals of the authors that permitted DNS to remain in
use (with some modifications) to this date when the internet has
experienced an unimagined exponential growth. And as far as we know, DNS
will continue providing reliable name resolution services for the internet
in the years to come.
Received on Mon Oct 03 2005 - 02:10:11 EDT

This archive was generated by hypermail 2.2.0 : Mon Oct 03 2005 - 09:02:41 EDT