Review - Development of the Domain Name System

The Domain Name System provides name to address translation for the
Internet. The architecture is organized into name servers, which are
repositories of information, and resolvers, which find information requested
by a client. The name space is a variable-depth hierarchical tree, where
administrative responsibilities may be delegated upward, maintaining a level
of flexibility common to the Internet.

This paper outlines several of the key considerations faced by early
designers of the DNS. Mockapetris et al. highlight the goals of the DNS,
including simplicity, extensibility, performance and maintainability. In
particular, zoning consideration and the delegation to subzones contributes
a great deal to the adoption of such a global system. Performance is also
discussed with a focus on caching and the time-to-live mechanism.

While the paper is an excellent recount of the early Internet, it is not a
scientific paper. The writers describe design goals and implementations
along with successes and failures of the design. Although some measurement
analysis is used as support, there is little scientific discussion and few
experimental results.

The Domain Name System is not without its faults. However, the designers
realized that in order to achieve a high level of adoption, the system would
need to be simple and flexible. As a result the design has lasted nearly a
quarter of a century and is an integral part of the Internet today.
Received on Mon Oct 03 2005 - 09:02:39 EDT