| Lesson 6 || Name servers |
| Objective || Describe the Three Classes of Name Servers. |
Three Classes of Name Servers
A name server
is a process that accepts queries into the DNS database.
It is also common usage to refer to the computer on which this process is running as a name server.
A name server process binds to the UDP
port and TCP
port 53 (the domain port
listed in /etc/services) and listens for incoming queries.
The UDP port is used for most database lookup queries (For example, what is the IP address of the machine named X?) while the TCP port is used for zone file transfers, which are described below.
Individual name servers play different roles in the overall operation of the DNS database.
Three important classes of name servers, primary servers, secondary servers, and root servers, are discussed below.
Every zone must have exactly one associated primary name server. This name server is the unique location that has the official or authoritative DNS information on its zone.
When an organization registers a new domain, it is required to supply the DNS administrators at the InterNIC with the IP address of an authoritative server for its zones.
The primary server is distinguished by the fact that the DNS data for its zone is stored on the server hosts local disk in a set of files called zone files.
In addition to its primary server, a zone may have one or more secondary servers. A secondary server provides an alternative source for information on the zone.
Secondary servers allow for load distribution (so that the authoritative server need not answer every query) and provide a backup (in case the authoritative server crashes or is down for maintenance.)
Administrators of DNS zones are required to supply at least one secondary server for each zone when they register their domain.
Secondary servers obtain their data from the authoritative server by periodically downloading a copy of the authoritative servers zone files.
This transfer is called a zone file transfer. Zone file transfers between name servers use TCP port 53.
A primary or secondary server for a zone is called authoritative because it obtains its information directly from data files describing the zone.
The root servers are the servers for the root domain. They are operated by the InterNIC.
At present, there are twelve root servers, named A.ROOT-SERVERS.NET, B.ROOT-SERVERS.NET, and so forth.
The root servers play a crucial role in binding together the DNS database through the
mechanism of recursion, which we will discuss in the next lesson.
User Datagram Protocol (UDP) is a procotol that sends data from one program to another using Internet Protocol (IP).
UDP is less reliable than TCP because it provides unverified transportation for individual messages.
Transmission Control Protocol (TCP) is a protocol that sends data from one program to another using Intenet Protocol (IP). TCP is connection-based and, therefore, more reliable than UDP.