Inside the Domain Name System

Name resolution is a key component in any network operating system (NOS) implementation. The capability of any one resource to locate other resources is the centerpiece of a functional network. Consequently, the name-resolution strategy chosen for a particular NOS must be robust and reliable, and it ideally will conform to industry standards.

Windows Server 2003 utilizes the Domain Name System (DNS) as its primary method of name resolution, and DNS is a vital component of any Active Directory implementation. Windows Server 2003's DNS implementation was designed to be compliant with the key Request for Comments (RFCs) that define the nature of how DNS should function. This makes it particularly beneficial for existing network implementations, as it allows Windows Server 2003 to interoperate with other types of RFC-compliant DNS implementations.

This chapter details the key components of DNS in general and provides an overview of Windows Server 2003's specific implementation of DNS. A particular emphasis is placed on the role of DNS in Active Directory and the way it fits in standard and nonstandard configurations. Step-by-step instructions outline how to install and configure specific DNS components on Windows Server 2003. In addition, troubleshooting DNS issues and specific Active Directory design scenarios help to give a hands-on approach to your understanding of DNS.

The Need for DNS

Computers and humans conceptualize in drastically different ways. In terms of understanding locations, humans are much better at grasping the concept of names rather than numbers. For example, most people think of cities by their names, not by their ZIP Codes. Computers, however, work in binary, and subsequently prefer to work with numbers. For example, computers at the post office translate the city and address names into specific ZIP Codes for that region, helping each letter reach its destination.

Name resolution for computer systems works in a similar way. A user-friendly name is translated into a computer-identifiable number. TCP/IP uses a number scheme that uniquely identifies each computer interface on a network by a series of numbers, such as 10.1.2.145, known as an IP address. Because most humans are not interested in memorizing several of these types of numbers, they must be easily resolvable into user-friendly names such as www.microsoft.com.

DNS, in its simplest form, provides for name resolution in a distributed fashion, with each server or set of servers controlling a specified zone and with entries for each resource called resource records (RRs) that indicate the location of a particular object.

A good analogy for DNS can be found in telephone books. Each city or metropolitan area (namespace) publishes a separate phone book (zone) that contains many listings (resource records) that map people's names to their phone numbers (IP addresses). This simple example illustrates the basic principle behind DNS. When you understand these basics, further drilling down into the specifics, especially with regard to Windows Server 2003's DNS, is possible.

DNS History

The Internet, as originally implemented, utilized a simple text file called a HOSTS file that contained a simple list of all servers on the Internet and their corresponding IP addresses. This file was copied manually from the master server to multiple secondary HOSTS servers. As more and more servers were added to the Internet, however, updating this file become unmanageable, and a new system became necessary.

In 1983, in direct response to this problem, the RFCs for the Domain Name System were drawn up, and this form of name resolution was implemented on a large scale across the Internet. Instead of a small number of static HOSTS files, DNS servers formed a hierarchical method of name resolution, in which servers resolved only a certain segment of hosts on the Internet and delegated requests that it did not manage. This allowed the number of records held in DNS to scale enormously, without a subsequent large performance decrease.

Microsoft developed its own implementation of DNS in Windows NT 4.0, which was based on the RFC standards on which DNS was founded. With the introduction of Windows 2000, Microsoft adopted DNS as the name-resolution strategy for Microsoft products. Older, legacy name-resolution systems such as WINS are slowly being phased out. Since that time, the DNS implementation used by Microsoft has evolved to include a number of key benefits that distinguish it from standard DNS implementations, such as those in other DNS implementationsfor example, Unix BIND. To understand these improvements, however, you first need a basic understanding of DNS functionality.